<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
	>

<channel>
	<title>Le blog groupe Reflect &#187; David Lafon</title>
	<atom:link href="http://www.groupereflect.net/blog/archives/author/david-lafon/feed" rel="self" type="application/rss+xml" />
	<link>http://www.groupereflect.net/blog</link>
	<description>attention marketing</description>
	<lastBuildDate>Thu, 02 Sep 2010 07:45:03 +0000</lastBuildDate>
	<generator>http://wordpress.org/?v=2.8.4</generator>
	<language>en</language>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
			<item>
		<title>Skipfish : outil d&#8217;audit de sécurité pour les applications web</title>
		<link>http://www.groupereflect.net/blog/archives/2010/03/skipfish-audit-securite-applications-web.html?parole_d_expert</link>
		<comments>http://www.groupereflect.net/blog/archives/2010/03/skipfish-audit-securite-applications-web.html?parole_d_expert#comments</comments>
		<pubDate>Wed, 24 Mar 2010 10:22:55 +0000</pubDate>
		<dc:creator>David Lafon</dc:creator>
				<category><![CDATA[Innovation]]></category>
		<category><![CDATA[Techno]]></category>
		<category><![CDATA[audit]]></category>
		<category><![CDATA[opensource]]></category>
		<category><![CDATA[php]]></category>
		<category><![CDATA[qualité]]></category>
		<category><![CDATA[securité]]></category>
		<category><![CDATA[test]]></category>

		<guid isPermaLink="false">http://www.groupereflect.net/blog/?p=2538?parole_d_expert</guid>
		<description><![CDATA[Chez groupeReflect, la cellulle Testing &#8211; Qualité utilise habituellement différents outils permettant de vérifier la sécurité des applications que nous développons. Parmi ceux-là, RATS et Nikto font partie de ceux que nous utilisons et qui font, également, partie des logiciels d&#8216;audit de sécurité d&#8217;applications web les plus usités sur le web. Cependant, cette hiérarchie pourrait [...]]]></description>
			<content:encoded><![CDATA[<p><img class="alignleft size-medium wp-image-2539" title="Sécurité des applications web - groupeReflect" src="http://www.groupereflect.net/blog/wp-content/uploads/2010/03/securite-182x300.png" alt="securite" width="164" height="270" />Chez <strong>groupeReflect</strong>, la cellulle Testing &#8211; Qualité utilise habituellement différents outils permettant de vérifier la <strong>sécurité</strong> des applications que nous développons. Parmi ceux-là, <a title="RATS - Rough Auditing Tool for Security" href="http://www.fortify.com/security-resources/rats.jsp" target="_blank">RATS</a> et <a title="Nikto2" href="http://cirt.net/nikto2">Nikto</a> font partie de ceux que nous utilisons et qui font, également, partie des logiciels d<strong>&#8216;audit de sécurité d&#8217;applications web</strong> les plus usités sur le web. Cependant, cette hiérarchie pourrait être modifiée avec l&#8217;arrivée d&#8217;un  nouveau venu : <strong><a title="skipfish - web application security scanner" href="http://code.google.com/p/skipfish/" target="_blank">skipfish</a></strong>.<span id="more-2538"></span></p>
<p>Publié sous licence Apache 2, ce soft est écrit  par Michal Zalewski, <em>hacker white hat</em>, employé du département Sécurité chez Google. Ecrit en C, <strong><a title="skipfish - web application security scanner" href="http://code.google.com/p/skipfish/" target="_blank">skipfish</a></strong> serait à même de simuler une foultitude d&#8217;attaques : injection SQL ou XML, Cross-Site Scripting (XSS), tentatives de connexion SSL&#8230; Avec la possibilité d&#8217;envoyer 500 requêtes à la seconde avec une configuration moyenne, sans pour autant prendre trop de ressource (un énorme avantage) , d&#8217;apprendre grâce à la reconnaissance heuristique des chemins et paramètres, de la gestion personnalisé de la pile HTTP, <strong><a title="skipfish - web application security  scanner" href="http://code.google.com/p/skipfish/" target="_blank">skipfish</a></strong> semble plus que prometteur. On pourra regretter qu&#8217;il ne prenne pas encore en compte l&#8217;ensemble des <a title="Web Application Security Scanner Evaluation Criteria Page history last edited by Brian Shura 18 hrs ago     Web Application Security Scanner     Evaluation Criteria     Version 1.0   Copyright © 2009 Web Application Security Consortium (http://www.webappsec.org)        Download this Document in PDF Format     Download the Evaluation Spreadsheet     List of Web Application Security Scanners        Table of Contents     Introduction Contributors  Contact Categories  Section 1 - Protocol Support  Section 2 - Authentication   Section 3 - Session Management  Section 4 - Crawling  Section 5 - Parsing  Section 6 - Testing  Section 7 - Command and Control  Section 8 - Reporting  A. Advice for Conducting a Scanner Evaluation  B. License    Introduction     Web Application Security Scanners are automated tools to test web applications for common security problems such as Cross-Site Scripting, SQL Injection, Directory Traversal, insecure configurations, and remote command execution vulnerabilities.  These tools crawl a web application and locate application layer vulnerabilities and weaknesses, either by manipulating HTTP messages or by inspecting them for suspicious attributes.     A large number of web application scanning tools are available, both commercial and open source.  Effective use of these tools is an important part of a thorough web application security assessment, and regular security scans are required to comply with security requirements such as section 6.6 of the Payment Card Industry Data Security Standard (PCI-DSS).     The Web Application Security Scanner Evaluation Criteria (WASSEC) is a set of guidelines to evaluate web application scanners on their ability to effectively test web applications and identify vulnerabilities.  It covers areas such as crawling, parsing, session handling, testing, and reporting.      The goal of the WASSEC is to create a vendor-neutral document to help guide web application security professionals during web application scanner evaluations.  This document provides a comprehensive list of features that should be considered when conducting a web application security scanner evaluation.  Different users will place varying levels of importance on each feature, and the WASSEC provides the user with the flexibility to take this comprehensive list of potential scanner features, narrow it down to a shorter list of features that are important to the user, assign weights to each feature, and conduct a formal evaluation to determine which scanning solution best meets the user's needs.     The aim of this document is not to define a list of requirements that all web application security scanners must provide in order to be considered a &quot;complete&quot; scanner, and evaluating specific products and providing the results of such an evaluation is outside the scope of the WASSEC project.  Instead, this project provides the tools and documentation to enable anyone to evaluate web application security scanners and choose the product that best fits their needs.  NIST Special Publication 500-269, &quot;Software Assurance Tools:  Web Application Security Scanner Functional Specification Version 1.0&quot;, contains minimal requirements for mandatory and optional web application scanner features.  This document can be found at https://samate.nist.gov.   Contributors     This document is the result of a team effort.  The following people have contributed their time and expertise to this project:         * Anurag Agarwal (Whitehat Security)     * Vijay Agarwal (Foundstone)     * Robert Auger (WASC)     * Emilio Casbas (S21sec)     * Leonardo Cavallari (NSRAV)     * Matthieu Estrade (Bee Ware)     * Romain Gaucher (Cigital, Inc.)     * Jeremiah Grossman (Whitehat Security)     * Robert Hansen (SecTheory)     * Amit Klein     * Chad Loder (Rapid7)     * Ken Pfeil (WestLB AG)     * Tyler Reguly (nCircle Network Security)     * Ivan Ristic (Breach Security)     * Ory Segal (IBM)     * Sheeraj Shah (Blueinfy Solutions Pvt. Ltd.)     * Chris Shiflett (OmniTI)     * Brian Shura (AppSec Consulting) [Project Leader]     * Tom Stripling (AppSec Consulting)     * Chris Sullo (Hewlett-Packard)    Contact     Participation in the Web Application Security Scanner Evaluation Criteria project is open to all.  If you have any questions about the evaluation criteria, please contact Brian Shura at bshura73@gmail.com.    Categories     Section 1 - Protocol Support     In order to test web applications, a scanner must support all communication protocols that are commonly used by web applications and intermediary network devices.  The underlying communication protocol used by web applications is the Hypertext Transfer Protocol (HTTP).     1.1  Transport Support  A web application scanner should support the following protocols:    1.1.1  HTTP 1.1:  This is the current version of HTTP and is the version that is most widely used by HTTP clients and servers.     1.1.2  HTTP 1.0:  This version of HTTP is still widely used, especially by proxy servers, and therefore should be supported by a web application scanner.     1.1.3  SSL/TLS:  Web applications that handle sensitive information employ Transport Layer Security to encrypt the data in transit.     1.1.4  HTTP Keep-Alive:  For performance reasons, a web application scanner should keep all connections open with the web server and re-use these connections for multiple HTTP requests unless instructed not to by the web server.     1.1.5  HTTP Compression:  Some web servers compress content before sending it to the client in order to improve performance.  Web application scanners should support HTTP compression for optimized performance and in order to properly parse compressed content encountered during a scan.   1.1.6  HTTP User Agent Configuration:  The ability of a web application scanner to mimic a specific browser can be very important.  Many web applications present different content depending on which browser the client is using.  A scanner should provide the user with the ability to configure a specific user agent string that is to be used during the course of the scan.     1.2  Proxy Support     It is not uncommon for a scanner to not have direct access to a website.  In this situation the scanner should allow the configuration of a proxy.  It should be noted that, while sometimes necessary, scanning a web application through proxy can result in false negatives and should be avoided if possible.     The following proxy capabilities should be supported by a scanner.     1.2.1  HTTP 1.0 proxy     1.2.2  HTTP 1.1 proxy     1.2.3  Socks 4 proxy     1.2.4  Socks 5 proxy     1.2.5  PAC File Support:  In some cases a scanner will need to follow special rules to determine which URLs to request through a proxy and which URLs to request directly.  A scanner should support the use of a proxy auto-config (PAC) file in order to properly handle these situations.   Section 2 - Authentication   This section covers support for standard or widely deployed authentication methods that web application scanners should support in order to effectively test applications that require authentication.           2.1  Authentication Schemes                    A web application security scanner should support the following authentication schemes:          2.1.1    Basic                   2.1.2    Digest                                    2.1.3    HTTP Negotiate (NTLM and Kerberos)                     2.1.4    HTML Form-based     2.1.4.1    Automated - The user supplies the scanner with a valid user id and password while configuring the scan.     2.1.4.2    Scripted - The scanner allows the user to script or record the login process prior to performing the scan.     2.1.4.3    Non-Automated - The scanner supports user intervention while the scan is being performed - for example, to solve a CAPTCHA or enter a value from a hardware token for a web application that has implemented two-factor authentication.                    2.1.5    Single Sign On (e.g. Cas, OpenID, AWS, AuthSub)                   2.1.6    Client SSL Certificates                   2.1.7    Custom implementations within the extensible HTTP authentication framework.   Section 3 - Session Management     During a web application security scan, it is crucial that scanners will maintain a “living” and valid session with the application at all times. A valid application session is mainly required for:         * Web crawling - In order to achieve the best possible coverage of the application, the scanner needs to be in a valid living session that will allow it to discover all possible web elements (e.g. parameters, cookies, forms, links, etc.).     * Test phase - Most application-layer security test cases require that the HTTP request sent by the scanner will be considered by the application as being &quot;in-session&quot;.  For example, testing an HTML form that requests bank account details, for SQL Injection.  If the scanner sends an HTTP request which does not have valid session tokens, the application will simply disregard this request and will probably redirect the scanner to the application's login page.    In order to maintain a valid session, the scanner needs to cope with all sorts of session management mechanisms.   3.1  Session Management Capabilities     Scanners should:     3.1.1  Understand that the application is asking it to start a new session, using a certain type of token as a method of uniquely identifying this session.     3.1.2  Perform a session token refresh, when instructed to do so by the application.     3.1.3  Detect that a currently held session was invalidated by the application (session expired).     3.1.4  Know how to start a new session and reacquire session tokens, in case the current session has expired.     3.2  Session Management Token Type Support     The most common types of session management tokens, which should be supported by a web application scanner, are:     3.2.1  HTTP Cookies (RFC 2965) - HTTP cookies are probably the most commonly used type of web application session tokens.  Web application scanners should be able to imitate the proper behavior of a web browser, such as accepting a request to set a new cookie and following a cookie value refresh instruction.     3.2.2  HTTP Parameters - Oftentimes, web applications use an HTTP parameter to track a web session.  There are several ways in which an application can instruct the web browser to use parameters as session tokens, for example, by embedding them in subsequent HTML links, or as a hidden HTML form parameter.  Web application scanners should be able to use HTTP parameters as session tokens.     3.2.3  HTTP URL Path - Some web applications embed session tokens inside the Path part of the URL.  For example:  http://example.com/app/{SESSION_TOKEN}/dir/file.aspx.  In such cases, a web application scanner should be able to cope with such a token format.  It is important to note that if the scanner is performing some sort of an application structure analysis during a scan, for example, in order to detect web directories, it will not treat such a session token as an actual part of the application structure, as this might generate irrelevant test cases.     3.3  Session Token Detection Configuration     Web application scanners should allow users to define the following session token configurations:     3.3.1  Automatic Session Token Detection and Value Refresh:  The scanner will attempt to detect session tokens on its own and will decide which tokens should be automatically tracked/refreshed during the scan.     3.3.2  Manual Session Token Configuration:  The user will define what denotes a session token, based on HTTP parameters, cookies, or any other type of configuration that is relevant (e.g. parse parts of the response and extract certain data from it, which serves as the session token value).     3.4  Session Token Refresh Policy     Session tokens are sometimes refreshed by the web application during a scanning session.  The scanner's session configuration should enable the user to define when, or during which phase of the scan, the session tokens should be refreshed.  The following configuration options should be provided by the scanner:     3.4.1  Fixed Session Token Value:  When a session token is marked to use a fixed value, this value will never change during the scan.     3.4.2  Login Process Provided Token Value:  Once the web application scanner logs into the application it will extract token values, which were issued as part of the login process, and will use them until it detects that the session was invalidated.     3.4.3  Dynamic Token Value:  The scanner will always use the most recent session token value, as supplied by the application at all times.  This means that if during the crawling or testing phase of the scan, a new value is detected, the scanner will stop and refresh all subsequent HTTP requests, with the most recent value.  This mode is extremely important for applications that make use of &quot;nonces&quot;, which dynamically change all the time.   Section 4 - Crawling     Crawling is the term used to describe the action taken by a program as it browses from page to page on a website.  The crawler will visit a starting page and parse the provided links, crawling to those pages.  This process will continue until some user-defined criteria is reached or the process is completed and there are no more links to crawl.     Crawling is essential to a web application security scan - it ensures that the scanner is aware of all linked pages that exist on the website.     Crawlers should be as configurable as possible, allowing the user to define a large number of criteria to ensure a thorough and efficient crawl.     4.1  Web Crawler Configuration     In regards to crawling, a web application scanner should:     4.1.1  Provide the user with the option to define a starting URL.  While '/' is commonly seen as the home URL of a website, this is not always the case.     4.1.2  Provide the user with the option to define additional hostnames (or IPs) in a list or a range.  A website may consist of several domains or subdomains (e.g. www.example.com, members.example.com, faq.example.com).  It is important that the user have the option of auditing these in a single scan.     4.1.3  Provide the user with the option to define exclusions for:     4.1.3.1  Specific hostnames (or IPs)     4.1.3.2  Specific URLs or URL patterns (regular expressions)     4.1.3.3  Specific file extensions     4.1.3.4  Specific parameters     4.1.4  Provide the user with the ability to limit redundant requests.  Large sites with catalogs or photo galleries will often have numerous redundant pages.  The capability to tune a crawler to limit the requests to these redundant pages should exist.     4.1.5  Provide the user with the option of supporting concurrent sessions.  The capability should exist to enable support for multiple sessions (logins) or limit crawling to a single session.     4.1.6  Provide the user with the ability to specify a request delay.  Network limitations may require that requests be throttled, while a high bandwidth network may not require this.  By setting a delay between requests, network traffic can be limited accordingly.     4.1.7  Provide the user with the option to define a maximum crawl depth.  The 'depth' of a website is a measurement of how many clicks it takes to reach a certain page from the starting page.  If you've clicked three links, then you're at a depth of three.  The user may want to limit their scan to pages that are two or three levels deep.     4.1.8  Provide the user with a method of training the crawler.  The process of training a crawler involves accessing a browser (internal or external to the scanning solution) and manually browsing a website.  As the website is browsed, valid pages and form inputs are recorded for the scanner to later use.   4.2  Web Crawler Functionality     During operation, a crawler should:     4.2.1  Identify newly discovered hostnames.  Websites commonly include links to other sites.  It is important to be able to identify when this is occurring so that the user can increase the scope of the crawl if necessary.     4.2.2  Support automated form submission.  Forms that accept varying data types are common on websites.  Quite often additional pages with new links exist behind these forms and the crawler should be able to access these additional pages.     4.2.3  Detect error pages and custom 404 responses.  The crawler should know when it has accessed an error page or found a custom 'Page Not Found' (404) response.     4.2.4  Redirect support.  The crawler should have the capability to:     4.2.4.1  Follow HTTP redirects.     Example  HTTP status codes 301, 302, 303, and 307     4.2.4.2  Follow Meta Refresh redirects.     Example  &lt;meta http-equiv=&quot;refresh&quot; content=&quot;1; URL=http://example.com/&quot;&gt;     4.2.4.3  Follow JavaScript redirects.     Example  &lt;script type=&quot;text/javascript&quot;&gt; window.location = &quot;http://www.example.com&quot;; &lt;/script&gt;   4.2.5  Identify and accept cookies.  During any visit to a website, many cookies may potentially be set.  The crawler should recognize these cookies, store them, and pass them back to the web server while crawling.     4.2.6  Support AJAX applications.  At a minimum, a crawler should be able to automatically submit XmlHTTPRequests that are found during the crawling process.    Section 5 - Parsing     In order to thoroughly scan a web application for security problems, a web application scanner must first map out the web application's structure and functionality.  The mapping process is done by the web crawler component, which makes use of different types of content parsers to extract information from web content.  This information may include URLs, HTML forms, HTML form parameters, HTML comments, and so forth.    5.1  Web Content Types     A scanner should be capable of parsing the following content types to extract information about the application's structure and functionality:    5.1.1  HTML:  HTML is the most elementary building block of the world-wide web.  Web application scanners must be able to fully parse and understand HTML content.     5.1.2  JavaScript:  JavaScript is probably the most common web client-side scripting language in use today.  Web application scanners should be able to parse JavaScript content to extract static and dynamic URLs.  This includes in-line Javascript embedded within &quot;script&quot; tags on HTML pages, script in standalone Javascript files (usually referred to from a &quot;script&quot; tag with a &quot;src&quot; attribute), and Javascript embedded within HTML tags in event handler attributes, such as an &quot;img&quot; tag with an &quot;onClick&quot; attribute.     5.1.3  VBScript:  Some web applications make use of VBScript as the language of choice for client-side scripting.  Although less popular than JavaScript (mainly due to the fact that it is currently not supported in many browsers other than Microsoft's Internet Explorer), VBScript content parsing should be supported by web application scanners.     5.1.4  XML:  XML content may appear in several scenarios when scanning web applications - for example, SOAP web services messages, Web Service Description Language (WSDL), XML Data Islands (XML Data embedded into an HTML page), WebDAV requests, XHTML, and so forth.  Web application scanners should be able to parse XML content.     5.1.5  Plaintext:  Some web applications may store information in plaintext files.  Web application scanners should be able to parse plaintext files and to extract relevant information.  A common example of a plaintext file that exists in most web applications and contains application structure information is the Robots.txt file.     5.1.6  ActiveX Objects:  Web applications may encapsulate client-side logic as ActiveX controls, which are downloaded and executed by the web browser.  Web application scanners should be able to extract application information from such controls.     5.1.7  Java Applets:  Similar to ActiveX objects, web developers may choose to encapsulate client-side logic as Java Applets.  Applets are small applications that are delivered as Java bytecode and executed on the client-side by the JVM (Java Virtual Machine).  Web application scanners should be able to extract application information from Java Applets.     5.1.8  Flash:  Adobe Flash is a very popular client-side platform for delivering multimedia content to web browsers.  Flash is often used to create interactive web applications with heavy animations and graphics, and was also adapted recently to support building RIAs (Rich Internet Applications).  Web application scanners should support extraction of content and information from applications that make use of Flash.     5.1.9  CSS:   Cascading style sheets (CSS) is a language widely used in web applications to describe the presentation of HTML documents.  Web application scanners should support the extraction of URLs from both inline CSS and external CSS files.    5.2  Character Encoding Support     Web applications often include content encoded in forms other than ASCII.  A web application scanner should be able to parse and understand content encoded in the following encoding types:   5.2.1  ISO-8859-1: Defined in RFC 2616 as the default encoding for HTTP content   5.2.2  UTF-7:  7-bit Unicode Transformation Format     5.2.3  UTF-8:  8-bit UCS/Unicode Transformation Format     5.2.4  UTF-16:  16-bit Unicode Transformation Format     5.3  Parser Tolerance     Often, web content may not conform to appropriate standards, either by mistake or due to various types of problems, ranging from bad developer habits to communication problems.  In such cases, it is crucial that content parsers will be able to cope with partial or non well-formed content and still be able to extract the relevant information from the application responses.  It's important that web applications be at least as robust in their HTML parsing as web browsers are.     5.4  Parser Customization     Since web technologies and standards evolve rapidly, web applications may include links and other types of information which is encapsulated in various ways not covered in section 4.1 of this document.  In order to support web applications that make use of such technologies and standards, it is recommended that web application scanners' parsers will allow user customization for link and content extraction.   Example - the following non-standard URL contains parameter names and values, which are concatenated by the string '::'. Each parameter name and value pair use the string '^^' to denote the equal sign:     http://www.some.site/appEntry?param1^^value1::param2^^value2...paramN^^valueN     In this scenario, the parser should be customizable to understand this non-standard URL format and properly parse the parameter names and values.     5.5  Extraction of Dynamic Content (Client-Side Logic Execution)     Due to the dynamic nature of client-side scripting languages, it is often not enough to be able to statically parse scripting languages such as JavaScript, VBScript, or even Flash applications in order to detect links and other relevant application information.  In such cases, web application scanners should be able to emulate user interaction with client-side logic in order to dynamically extract this information.   Section 6 - Testing     Testing an application for vulnerabilities is the core functionality of a web application security scanner.  This section lists the types of vulnerabilities that a web application scanner should be capable of detecting, as well as the testing-related configuration and customization options that a scanner should provide.   6.1  Testing Configuration   It is sometimes important to prevent the web application scanner from testing a given part of a web application. A scanner should provide the ability to reduce its visibility of the web application based on different criteria.     6.1.1  Host names or IPs - The tester should be able to specify specific host names or IP addresses to be ignored by the web application scanner during the test phase.  Note that the scanner might be able to crawl and parse these host names and IP addresses for reasons such as logging in and data gathering, but should not test them.     6.1.2  URL patterns - Using patterns (e.g., regular expression, etc.), the tester should be able to specify which URLs should not be tested by the web application scanner.     6.1.3  File extensions - The user should have the ability to specify which extensions to exclude from testing.     6.1.4  Parameters - The user should have the ability to specify specific input parameters to exclude from testing.     6.1.5  Cookies - It is sometimes easier to specify which part of the tested web application the scanner should ignore based on cookie specific values.  The tester should then be able to specify one or many cookie values that would prevent the scanner from testing the web page.     6.1.6  HTTP headers - The user should have the ability to configure the scanner to exclude certain pages from testing based on specific HTTP response headers.     6.2  Testing Capabilities     A web application scanner is looking for two major types of security problems: vulnerabilities and architectural weaknesses.  The following list of problems to test is mostly extracted from the WASC Threat Classification 2.0.     6.2.1  Authentication                          6.2.1.1 Brute Force                                    6.2.1.1.1 Lack of Account Lockout                                    6.2.1.1.2 Different Login Failure Message For Valid vs. Invalid User names                          6.2.1.2 Insufficient Authentication                          6.2.1.3 Weak Password Recovery Validation                          6.2.1.4 Lack of SSL On Login Pages                          6.2.1.5 Auto-complete Not Disabled On Password Parameters    6.2.2  Authorization           6.2.2.1 Credential/Session Prediction                     6.2.2.1.1 Sequential Session Token                       6.2.2.1.2 Non-Random Session Token           6.2.2.2 Insufficient Authorization                     6.2.2.2.1 Ability To Forcefully Browse To &quot;Logged-In&quot; URL Without Logging In                     6.2.2.2.2 Ability To Forcefully Browse To High-Privilege URL While Logged In Under Low-Privilege Account                     6.2.2.2.3 HTTP Verb Tampering           6.2.2.3 Insufficient Session Expiration                          6.2.2.4 Session Fixation                                    6.2.2.4.1 Failure To Generate New Session ID After Login                                    6.2.2.4.2 Permissive Session Management                          6.2.2.5 Session Weaknesses                                    6.2.2.5.1 Session Token Passed In URL                                    6.2.2.5.2 Session Cookie Not Set With Secure Attribute                                    6.2.2.5.3 Session Cookie Not Set With HTTPOnly Attribute                                    6.2.2.5.4 Session Cookie Not Sufficiently Random                                    6.2.2.5.5 Site Does Not Force SSL Connection                                    6.2.2.5.6 Site Uses SSL But References Insecure Objects                                    6.2.2.5.7 Site Supports Weak SSL Ciphers         6.2.3  Client-side Attacks                          6.2.3.1 Content Spoofing                          6.2.3.2 Cross-Site Scripting                                    6.2.3.2.1 Reflected Cross-Site Scripting                                    6.2.3.2.2 Persistent Cross-Site Scripting                                    6.2.3.2.3 DOM-Based Cross-Site Scripting                          6.2.3.3 Cross-Frame Scripting                          6.2.3.4 HTML Injection                          6.2.3.5 Cross-Site Request Forgery                          6.2.3.6 Flash-Related Attacks                                    6.2.3.6.1 Cross-Site Flashing                                    6.2.3.6.2 Cross-Site Scripting Through Flash                                    6.2.3.6.3 Phishing/URL Redirection Through Flash                                    6.2.3.6.4 Open Cross-Domain Policy         6.2.4  Command Execution                          6.2.4.1 Format String Attack                          6.2.4.2 LDAP Injection                          6.2.4.3 OS Command Injection                          6.2.4.4 SQL Injection                                    6.2.4.4.1 Blind SQL Injection                          6.2.4.5 SSI Injection                          6.2.4.6 XPath Injection                          6.2.4.7 HTTP Header Injection / Response Splitting                          6.2.4.8 Remote File Includes                          6.2.4.9 Local File Includes                          6.2.4.10 Potential Malicious File Uploads         6.2.5  Information Disclosure                          6.2.5.1  Directory Indexing                          6.2.5.2  Information Leakage                                    6.2.5.2.1  Sensitive Information In Code Comments                                    6.2.5.2.2  Detailed Application Error Messages                                    6.2.5.2.3  Backup Files (home.old, home.bak, etc.)                                    6.2.5.2.4  Include File Source Code Disclosure                          6.2.5.3  Path Traversal                          6.2.5.4  Predictable Resource Location                          6.2.5.5  Insecure HTTP Methods Enabled                          6.2.5.6  WebDAV Enabled                          6.2.5.7  Default Web Server Files                          6.2.5.8  Testing and Diagnostics Pages (test.asp, phpinfo.htm, etc.)                          6.2.5.9  Front Page Extensions Enabled                          6.2.5.10  Internal IP Address Disclosure     6.3  Testing Customization     Even if they use the same basic technologies and communication protocols,  web applications contain specific vulnerabilities that require very precise payloads to be enabled.  A web application scanner should:     6.3.1  Allow the user to modify existing tests.  It should be possible to either modify the attack steps (payload or algorithm) and/or the detection criteria.     6.3.2  Allow the user to create new customized tests.  The web application scanner might have a specific API to allow the creation of new tests, or simply allow the user to add new tests based on templates, regular expressions, etc.     6.4  Test Policy     Most web application scanners have a large number of built-in tests, but in many situations only a subset of these tests will be needed.  A web application scanner should allow the user to create personalized test policies that specify which tests to include in a scan.   Section 7 - Command and Control            The Command and Control capabilities of web application scanners can have a significant influence on usability and therefore are an important aspect to consider when conducting an evaluation.  The types of Command and Control features most valued by an end user will vary based on the user's situation - some of the following features will be important to a large enterprise with many users and web applications to scan, but will not necessarily apply to a small company with a single user looking for an effective, low-cost scanning solution.                       7.1  Scan Control Capabilities                     The following control capabilities should be considered during a web applicaton scanner evaluation:                    7.1.1  Ability to schedule scans.  Often the owners of web applications will only allow the applications to be scanned during specified time windows.  Therefore, it is important that a web application scanner provide the user with the ability to schedule scan.  This scheduling should include a start time as well as a maximum scan duration, after which the scan will be paused if it has not yet completed.                   7.1.2  Ability to pause and resume scans.  A scanner should provide the operator with the ability to pause a scan and then resume the scan at a later time from the point at which it was paused.                                        7.1.3  Ability to view real-time status of running scans.  A scanner should provide the operator with a way of viewing the real-time status of                     running scans.  This status could include information such as which tests are currently being run and the scan completion percentage.                   7.1.4  Ability to define re-usable scan configuration templates.  Often a significant amount of time and effort is involved in optimally configuring a web application scanner for a particular application.  A scanner should provide the user with the ability to save a scan's configuration so that it can be re-used for later scans.                   7.1.5  Ability to run multiple scans simultaneously.  Organizations that have many web applications to scan will find that the ability to run multiple scans at the same time is a desirable feature.                   7.1.6  Ability to support multiple users.  For organizations that plan to roll out a web application scanning solution to many users, the scanner's ability to support multiple users should be considered.  For example, some scanning solutions will need to be installed and run on each user's workstation.  Others provide a centralized web-based management interface that allows multiple users to configure and schedule scans and view scan results.                   7.1.7  Ability to support remote/distributed scanning.  The ability to support remote or distributed scanning is very important for decentralized environments where web applications are located behind firewalls or must be accessed over VPNs or slow links.  To effectively scan web applications in these environments, a web application scanner should provide the ability to deploy scanning &quot;agents&quot; inside an organization's network, which can be securely controlled by a remote operator.                       7.2  Control Interfaces Provided                     Web applicaton scanners can provide a variety of interfaces for the user to control the scanner:                7.2.1  Client Application with GUI.  Some scanners are designed to be installed on each user's workstation and provide a graphical user interface for configuring and controlling scans.                   7.2.2  Command Line Interface.  Some scanners provide a command line interface, with command options and/or configuration files used to input the scan settings.                    7.2.3  Web-Based Interface.  Some scanning solutions provide a web-based command and control interface.  This is especially common in systems that support multiple users.                               7.3  Extensibility and Interoperability           The ability to extend the functionality a web application scanner and integrate it with an organization's defect tracking systems can be important considerations for some users.  The following are extensability and interoperability features that can be provided by a web application scanning solution:                    7.3.1  Scan API.  Advanced users of web application scanners will sometimes find it useful to write their own code for controlling the scanner, executing custom tests, extracting data for custom reports, etc.  For these types of users, the existence of an Application Programming Interface (API) should be considered.  A scan API should be well-documented and supported.  Ideally, example code which shows how to use the API should be available.                   7.3.2  Ability to integrate with common bug-tracking systems.  Organizations often use defect tracking systems to track the status of web application defects.  A web application scanner should provide the ability to send vulnerabiliy information to bug-tracking systems.                 Section 8 - Reporting     In order for scanning results to be viewed outside of the tool's interface, web application scanners should be able to generate reports of each scan.  Because reports are often used by different groups within an organization, scanners should provide the ability to customize the format and information included in their reports.     8.1  Types of Reports              Although the specific reporting options may vary, scanners should provide the following types of reports:     8.1.1  Executive Summary     An executive summary provides a concise picture of the scan results.  This report allows a reader to able to determine high-level results at a glance.     8.1.2  Technical Detail Report     Scanners must be able to provide all technical information required for readers to reproduce the identified issues.  This should include:     8.1.2.1  The ability to include full request and response data     8.1.2.2  The ability to include a list of all hosts and URLs included in the scan     8.1.3  Delta Report     Scanners should provide the ability to compare results from two or more scans and show differences or trends over time.     8.1.4  Compliance Report     Scanners should provide a report format that allows organizations to quickly determine whether they are in violation of regulatory requirements or other standards.  These reporting capabilities should be considered if certain regulations are important to the organization.  The following list provides some potentially applicable standards:     8.1.4.1   OWASP Top 10     8.1.4.2   WASC Threat Classification     8.1.4.3   SANS Top 20     8.1.4.4   Sarbanes-Oxley (SOX)     8.1.4.5   Payment Card Industry Data Security Standard (PCI DSS)     8.1.4.6   Health Insurance Portability and Accountability Act (HIPAA)     8.1.4.7   Gramm-Leach-Bliley Act (GLBA)     8.1.4.8   NIST 800-53     8.1.4.9   Federal Information Security Management Act (FISMA)     8.1.4.10  Personal Information Protection and Electronic Documents Act (PIPEDA)     8.1.4.11  Basel II     8.2  Advisories For Each Unique Vulnerability Type     Scanners should have the capability to produce advisories for each unique vulnerability type they identify.  These advisories should include the following information:     8.2.1  Vulnerability description     8.2.2  CVE or CWE ID     8.2.3  Severity level     8.2.4  CVSS version 2 Score     8.2.5  Remediation guidance     8.2.6  Remediation code example(s)     8.3  Report Customization               Scanners should provide the ability to customize their reports, including the following:     8.3.1  Ability to add custom notes to vulnerabilities, which will be included in any generated reports.     8.3.2  Ability to mark vulnerabilities as false positives and remove them from the report.  Note that it should be possible for the scanner to list and/or report on all of the &quot;excluded&quot; vulnerability findings.  Ideally, the scanner should document who marked the vulnerability as a false positive, when it was marked, and the justification.        8.3.3  Ability to adjust the risk level of vulnerabilities, including:     8.3.3.1  Base, Temporal, and Environmental metrics that affect CVSS score     8.3.3.2  Severity level or other risk quantifiers     8.3.4  Ability to identify and report vulnerabilities according to their content location, which could be: URL, Portlet Name, Page Title, or User Defined (e.g. a regular expression pattern in the response).     8.3.5.: Ability to include customizations such as addition of a company logo or modification of report footer and header.     8.4  Report Format     Scanners should provide the capability to generate reports in both human and machine-readable formats, including:     8.4.1  PDF     8.4.2  HTML     8.4.3  XML     8.5  Vendor Feedback     Scanners should provide the ability to automatically report false positives or other types of feedback to the scanner vendor to help improve the quality of future versions of the product.  This information should be encrypted in transit and handled securely by the vendor.    A. Advice for Conducting a Scanner Evaluation     Web application scanners are complex pieces of software that can be a challenge to evaluate.  However, conducting an evaluation is a valuable learning experience and will help you identify the tool or tools that best meet your needs.  This appendix provides advice for conducting a successful evaluation and covers each phase of the evaluation process.     Preparation     Solid preparation is essential for conducting a successful evaluation.  During this phase, the following steps should occur:        1. Determine which of the criteria detailed in this document are most important to you as a user.  Some features will be essential &quot;must haves&quot;, while others will be less important and some will be of no value for your particular situation.  A future release of the WASSEC will include an evaluation spreadsheet that will allow you to adjust the weight of each criterion according to your specific needs.    2. There are many non-technical items that could be considered when evaluating a security scanner.  The importance of these criteria will vary depending on your environment and the intent of your evaluation.  For example, the importance of cost may be different between large and small organizations.  If desired, these items may be added to your evaluation spreadsheet and weighted according to your needs.  Some items to consider include:           * Purchase cost           * Ongoing support cost           * Additional cost (hardware, etc.)           * Ease of operation for user (e.g. developer, security analyst, QA analyst)           * Quality of documentation           * Support availability (phone, web, user community, etc.)           * Availability of training           * Product and/or testing capability update frequency           * Licensing restrictions    3. Decide which scanners will be in-scope for the evaluation.  Keep in mind that conducting a thorough evaluation can be time-consuming, therefore it's best to use your &quot;must have&quot; criteria to limit your scope to a short list of candidates to evaluate.    4. Obtain the latest version of each scanner you're going to evaluate.  For commercial scanners, vendors will normally provide a fully-functional copy of the software for a limited trial period if you're doing a private evaluation.     5. Decide which web applications will be scanned during the evaluation.  Some tips:           * Using a well-known vulnerable application such as WebGoat is unlikely to provide realistic results because the commercial scanners have most likely been optimized for these applications.  It's best to choose real-world applications.           * Select web applications that use a variety of technologies that are representative of your needs.  For example, if your company owns a mix of ASP.Net and Java applications, make sure to include both of these technologies in the applications you choose to include in the evaluation.           * Select web applications that employ a variety of design patterns that are representative of your needs.  For example, if you include a blogging website as Application A, a shopping website as Application B, and a site that requires users to log in and provides multiple levels of access as Application C, you'll learn much more from your evaluation than if you choose three applications of the same type.    6. Prepare the applications for scanning.  It's best to use a non-Production environment, especially if the applications haven't been scanned before. If you do choose to scan a Production environment during your evaluation, ensure that all databases used by the application have been backed up and Production Support personnel have been notified of your plans.    7. If you intend to release the results of your evaluation to the public, it is important to record the details of your application setup (for example, application version numbers, web server type and version, database type and version, and operating system) so that others can reproduce your results.     Scanning     Use the products to scan each of your in-scope web applications.  Since most scanners are very configurable, it will often be desirable to run multiple scans against each application using a variety of configurations.  For example, you'll likely see very different results from a &quot;point and shoot&quot; scan with minimal configuration versus a scan that is configured to log into the web application and manually trained on how to perform all significant transactions.     While configuring and running these scans, evaluate the products based on the criteria you rated as important during the Preparation phase.  For example, if Ease of Use is an important criterion in your evaluation, give each product a rating for this criterion.     Results Analysis           For each scanner covered in the evaluation, add up the weighted scores for each section of the WASSEC as well as the scores for any additional criteria you've included, such as cost, ease of use, etc.  If you were unsure how to rate certain items, contacting the scanner developer for additional information on specific aspects of the tool can help to fill in the gaps and ensure a thorough and fair evaluation.  After completing the evaluation, you'll be armed with the information needed to make an informed decision on which scanning solution best meets your needs.   B. License     This work is licensed under the Creative Commons Attribution License.  To view a copy of this license, visit http://creativecommons.org/licenses/by/2.5/ or  send a letter to: Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA.                           Web Application Security Scanner Evaluation Criteria 	 Page Tools Insert links  Insert links to other pages or uploaded files. Pages Images and files Insert a link to a new page     1.       Loading...     1. No images or files uploaded yet.  Insert image from URL  Tip: To turn text into a link, highlight the text, then click on a page or file from the list above. Comments (0)  You don't have permission to comment on this page. Printable version" href="http://projects.webappsec.org/Web-Application-Security-Scanner-Evaluation-Criteria" target="_blank">critères d&#8217;évaluation de sécurité</a> listés par le <a title="Web Application Security Consortium" href="http://www.webappsec.org/" target="_blank">Web Application Security Consortium</a> mais <strong><a title="skipfish - web application security  scanner" href="http://code.google.com/p/skipfish/" target="_blank">skipfish</a></strong> a emprunté le bon chemin et viendra renforcer, après évaluation et confirmation, les batteries de tests que nous lançons sur  nos applications.</p>
]]></content:encoded>
			<wfw:commentRss>http://www.groupereflect.net/blog/archives/2010/03/skipfish-audit-securite-applications-web.html?parole_d_expert/feed</wfw:commentRss>
		<slash:comments>4</slash:comments>
		</item>
		<item>
		<title>De retour de Paris web</title>
		<link>http://www.groupereflect.net/blog/archives/2009/10/de-retour-de-paris-web.html?parole_d_expert</link>
		<comments>http://www.groupereflect.net/blog/archives/2009/10/de-retour-de-paris-web.html?parole_d_expert#comments</comments>
		<pubDate>Mon, 19 Oct 2009 07:40:52 +0000</pubDate>
		<dc:creator>David Lafon</dc:creator>
				<category><![CDATA[Techno]]></category>
		<category><![CDATA[Utilisabilité]]></category>
		<category><![CDATA[Blog]]></category>
		<category><![CDATA[ergonomie]]></category>
		<category><![CDATA[groupereflect]]></category>
		<category><![CDATA[Paris web]]></category>
		<category><![CDATA[qualité]]></category>
		<category><![CDATA[roi]]></category>
		<category><![CDATA[Standards]]></category>
		<category><![CDATA[TIC]]></category>
		<category><![CDATA[W3C]]></category>
		<category><![CDATA[web]]></category>

		<guid isPermaLink="false">http://www.groupereflect.net/blog/?p=2115?parole_d_expert</guid>
		<description><![CDATA[Paris Web 2009 est terminé&#8230; Vive Paris Web 2010 ! Les 2 jours de conférence ont été un énorme condensé de qualité, de bonnes pratiques, de standards et de bonne humeur. Tout d&#8217;abord, je souhaite donner un grand coup de chapeau à toute l&#8217;équipe de Paris Web. Ils ont abattu un travail énorme pour nous [...]]]></description>
			<content:encoded><![CDATA[<p><a href="http://www.paris-web.fr/2009/"><img class="alignleft size-full wp-image-2118" src="http://www.groupereflect.net/blog/wp-content/uploads/2009/10/parisWeb_125.png" alt="Paris Web 2009" width="230" height="125" /></a><strong><a title="Paris Web 2009" href="http://www.paris-web.fr/2009/">Paris Web 2009</a></strong> est terminé&#8230; Vive Paris Web 2010 ! Les 2 jours de conférence ont été un énorme condensé de qualité, de bonnes pratiques, de standards et de bonne humeur. Tout d&#8217;abord, je souhaite donner un grand coup de chapeau à toute l&#8217;équipe de Paris Web. Ils ont abattu un travail énorme pour nous assurer un cycle de conférences d&#8217;une exceptionnelle qualité. Jugez plutôt, pas moins de 43 orateurs se sont relayés dans 2 amphithéâtres et 16 ateliers. Et pas des moindres :</p>
<ul>
<li>l&#8217;inénarrable et inébranlable (il a été orateur à tous les Paris Web depuis leur création !) <a title="Daniel Glazman : Paris Web 2009" href="http://www.glazman.org/weblog/dotclear/index.php?post/2009/10/08/Ma-presentation-a-ParisWeb-2009">Daniel Glazman</a>, <em>co-chairman</em> du <em>Working Group CSS </em>au W3C</li>
<li>la pétulante <a title="Molly E. Holzschlag" href="http://molly.com/">Molly E. Holzschlag</a> d&#8217;Opera Software et accessoirement éminent membre du <em>Working Group CSS</em> et expert au<em> Working Group HTML<br />
</em></li>
<li><a title="Standblog de Tristan Nitot" href="http://standblog.org/">Tristan Nitot</a>, jamais avare de bons mots, président de Mozilla Europe</li>
<li>le globe-trotter, <a title="Karl Dusbost - LaGrange.net" href="http://www.la-grange.net/">Karl Dubost</a>, ancien employé du W3C et Directeur technique chez <a title="Phéromone" href="http://www.pheromone.ca/">Phéromone</a></li>
<li>le réjouissant <a title="Charles McCathieNevile - Chaals" href="http://my.opera.com/chaals/blog/">Charles McCathieNevile</a>, <em>chief Standards Officer</em> chez Opera Software</li>
<li>la charmante <a title="Ergolab - Amélie Boucher" href="http://www.ergolab.net/">Amélie Boucher</a>, ergonome reconnue pour ses travaux et son très bon livre sur l&#8217;ergonomie des sites web</li>
<li><a title="Eric Daspet" href="http://eric.daspet.name/">Eric Daspet</a>, responsable innovation et consultant chez SQLi mais surtout évangéliste d&#8217;un web ouvert, expert francophone en PHP et un des pères fondateurs de Paris Web</li>
<li>et j&#8217;en oublie mais la liste complète se trouve sur le site de <a title="Paris web 2009 - les orateurs" href="http://www.paris-web.fr/2009/-orateurs-">Paris Web</a></li>
</ul>
<p>Finalement, tout le beau monde de l&#8217;Open Web était présent. Alors que retenir de ces 2 jours de conférence ?</p>
<ol>
<li>Parmi les infos exclusives divulguées, une est particulièrement ressortie et a notamment marqué les esprits : <strong>la prédiction de Daniel Glazman</strong> ! Selon lui, et je suis tout enclin à le croire, <strong>IE 6 sera devenu une part négligeable du marché des navigateurs en mai 2010 et aura complètement disparu en mai 2011</strong>.  Je ne sais pas sur quels chiffres il s&#8217;appuie mais je le crois, la courbe d&#8217;utilisation IE6 va dans ce sens. Cela rejoint les estimations que j&#8217;avais faites et sur lesquels nous basons nos cadres d&#8217;évaluation. Au final, est-il encore opportun de perdre du temps, de l&#8217;énergie et donc de l&#8217;argent sur la compatibilité multi-navigateurs en y incluant IE6. Je suis un fervent défenseur de l&#8217;arrêt du support d&#8217;IE6 (et au moins dans un premier temps du support partiel que j&#8217;appelle souvent &laquo;&nbsp;dégradation propre&nbsp;&raquo;) et ce pour 2 raisons :
<ul>
<li>les utilisateurs d&#8217;IE6 le sont sur leur lieu de travail donc pas forcément dans la cible directe d&#8217;une application web (sauf cas des intranets antédiluviens ou d&#8217;application de type B2B particulière) et finalement, dans notre métier, nous avons une responsabilité importante vis à vis des utilisateurs (je ne parle pas que de nos clients mais bien de l&#8217;ensemble des internautes), à savoir, engager un processus pédagogique d&#8217;évangélisation et de sensibilisation à la qualité finale de l&#8217;expérience utilisateur qu&#8217;ils sont en train de vivre. Rendre compatible une application web avec IE6 est une contrainte forte à la créativité et à l&#8217;innovation, en plus d&#8217;être une aberration commerciale !</li>
<li>sinon, il s&#8217;agit d&#8217;un choix délibéré de leur part et il convient de se poser la question de savoir si ils font vraiment partie de la cible de notre projet et si par conséquent il est opportun de dépenser une fortune pour rendre notre application compatible avec ce navigateur en sachant que le ROI n&#8217;est absolument pas assuré. Cette cible et son coût d&#8217;acquisition intéresse-t-elle vraiment notre client ? Je ne crois pas tant les coûts de mise en compatibilité deviennent exorbitants avec la complexification des applications web.</li>
</ul>
</li>
<li>Nous avons un devoir d&#8217;éducation, au sens pédagogique, envers les internautes. Il convient de leur fournir un produit de qualité, respectueux des standards (et j&#8217;y tiens), accessible et interopérable.<br />
Mais avoir une action pédagogique ne signifie pas simplement fournir un produit aux standards mais également orienter l&#8217;utilisateur final vers une interface qui lui offre la bonne (sinon la meilleure) expérience utilisateur. On en revient à l&#8217;infatigable débat sur IE6 qui gangrène les développeurs web depuis 8 ans et qui continue à brider l&#8217;innovation technologique des applications web :  le lissage par le bas ! Un des leitmotive de ce <strong><a title="Paris Web 2009" href="http://www.paris-web.fr/2009/">Paris Web 2009</a></strong> fût donc l&#8217;abandon pur et simple d&#8217;IE6. Oui&#8230; facile à dire ! Cependant la réalité du marché est celle qui nous fait vivre, on ne peut pas encore dire à nos clients qu&#8217;on ne supportera pas ou peu (mode dégradé)  IE6 dans l&#8217;application qu&#8217;on est en train de lui construire sans qu&#8217;il y ait des grincements de dents.</li>
<li>L&#8217;avenir des standards nous promet de belles choses. La conférence de Tristan Nitot et de Paul Rouget fût à ce point magistrale que toute l&#8217;assistance en fût bluffée. En effet, hormis le sketch bien huilé (et parfois comique) de 2 geeks qui se respectent et s&#8217;admirent mutuellement, la démonstration technologique HTML5/CSS3 (certes sur un Firefox 3.7 en pré-alpha) a scotché l&#8217;ensemble de l&#8217;assemblée : <a title="Mozbox: Some new demos" href="http://blog.mozbox.org/post/2009/10/12/Some-new-demos">détection de l&#8217;accéléromètre dans le navigateur, filtres CSS sur des vidéos et sur du canvas 3D</a>&#8230; Ces technologies laissent augurer d&#8217;un web plus puissant, plus interactif et plus ergonomique&#8230; que du bonheur !</li>
<li>La conférence d&#8217;Amélie Boucher a également été source de bonnes pratiques en ergonomie des sites e-commerce.  Elle a rejoint en divers points <a title="Thierry Rousseau : Eviter les abandons de panier" href="http://www.thierryrousseau.net/eviter-les-abandons-de-panier/">l&#8217;avis de Thierry Rousseau</a> sur l &#8216;amélioration des taux de conversion. Amélie est une fervente adepte (et je la comprends) de l&#8217;A/B testing pour s&#8217;adapter pleinement à sa cible.</li>
<li>De manière générale, j&#8217;ai été satisfait de l&#8217;ensemble des conférences dans la mesures où les pratiques et technologies citées  sont des chemins que nous abordons quotidiennement chez groupeReflect et sur lesquelles nous cherchons à avoir une vision lointaine pour nos clients.</li>
</ol>
<p>Pour conclure : une manifestation excellente en tous points. Par des professionnels, pour des professionnels. Le seul regret que je puisse y voir, c&#8217;est qu&#8217;il faille attendre une année pour y retourner et recharger les batteries !</p>
]]></content:encoded>
			<wfw:commentRss>http://www.groupereflect.net/blog/archives/2009/10/de-retour-de-paris-web.html?parole_d_expert/feed</wfw:commentRss>
		<slash:comments>1</slash:comments>
		</item>
		<item>
		<title>Paris Web 2009, nous y serons !</title>
		<link>http://www.groupereflect.net/blog/archives/2009/10/paris-web-2009-nous-y-serons.html?parole_d_expert</link>
		<comments>http://www.groupereflect.net/blog/archives/2009/10/paris-web-2009-nous-y-serons.html?parole_d_expert#comments</comments>
		<pubDate>Fri, 02 Oct 2009 12:47:30 +0000</pubDate>
		<dc:creator>David Lafon</dc:creator>
				<category><![CDATA[Evènements]]></category>
		<category><![CDATA[Utilisabilité]]></category>
		<category><![CDATA[Design]]></category>
		<category><![CDATA[ergonomie]]></category>
		<category><![CDATA[qualité]]></category>
		<category><![CDATA[Standards]]></category>
		<category><![CDATA[web]]></category>
		<category><![CDATA[Web 2]]></category>

		<guid isPermaLink="false">http://www.groupereflect.net/blog/?p=2046?parole_d_expert</guid>
		<description><![CDATA[
Pour ceux qui ne connaissent pas, Paris Web est une manifestation dédiée au monde du… web. Il s’agit de 2 jours de conférence  par des orateurs qui figurent parmi les sommités du web (Daniel Glazman, Karl Dubost, Tristan Nitot, Amélie Boucher ou encore Molly Holzschlag, etc.) et, le samedi, une journée d’ateliers également proposés [...]]]></description>
			<content:encoded><![CDATA[<p><a title="Paris-Web 2009" href="http://www.paris-web.fr/"><img class="alignleft" src="http://www.paris-web.fr/2009/IMG/gif/pw2009-120x90-2.gif" alt="Paris-web 2009 — 8–9–10 octobre — on y sera !" width="120" height="90" /></a><br />
Pour ceux qui ne connaissent pas, <strong><a title="Paris Web 2009 - qualité, standards, accessibilité" href="http://www.paris-web.fr/">Paris Web</a></strong> est une manifestation dédiée au monde du… web. Il s’agit de 2 jours de conférence  par des orateurs qui figurent parmi les sommités du web (<a title="&lt;Glazblog/&gt;, le blog de Daniel Glazman" href="http://glazman.org/weblog/">Daniel Glazman</a>, <a title="La-Grange.Net" href="http://www.la-grange.net/">Karl Dubost</a>, <a title="Standblog" href="http://standblog.org/">Tristan Nitot</a>, <a title="Ergolab" href="http://www.ergolab.net/">Amélie Boucher</a> ou encore <a title="Molly Holzschlag" href="http://molly.com/">Molly Holzschlag</a>, etc.) et, le samedi, une journée d’ateliers également proposés par des ténors du web (<a title="Laurent Denis - Temesis" href="http://www.temesis.com/a-propos/equipe-temesis/laurent-denis-expert-accessibilite.html">Laurent Denis</a>, <a title="Aurélien Levy - Temesis" href="http://www.temesis.com/a-propos/equipe-temesis/aurelien-levy-expert-accessibilite.html">Aurélien Levy</a>, <a title="Webatou - Monique Brunel" href="http://blog.webatou.info/">Monique Brunel</a>&#8230;) . Si vous êtes un professionnel du web et que vous êtes particulièrement sensible aux standards, à l’accessibilité, au design, à l’utilisabilité, aux performances et à la sécurité des applications web, nous aurons probablement l’occasion de nous croiser et d&#8217;échanger et débattre ensemble .</p>
<p>Au plaisir de vous y rencontrer !</p>
]]></content:encoded>
			<wfw:commentRss>http://www.groupereflect.net/blog/archives/2009/10/paris-web-2009-nous-y-serons.html?parole_d_expert/feed</wfw:commentRss>
		<slash:comments>1</slash:comments>
		</item>
		<item>
		<title>Facebook et les standards W3C</title>
		<link>http://www.groupereflect.net/blog/archives/2009/08/facebook_et_les.html?parole_d_expert</link>
		<comments>http://www.groupereflect.net/blog/archives/2009/08/facebook_et_les.html?parole_d_expert#comments</comments>
		<pubDate>Mon, 24 Aug 2009 09:48:38 +0000</pubDate>
		<dc:creator>David Lafon</dc:creator>
				<category><![CDATA[Techno]]></category>
		<category><![CDATA[facebook]]></category>
		<category><![CDATA[Standards]]></category>
		<category><![CDATA[TIC]]></category>
		<category><![CDATA[W3C]]></category>

		<guid isPermaLink="false">http://welkom.groupereflect.net/gr2009/blog/archives/2009/08/facebook-et-les-standards-w3c.html?author=-3</guid>
		<description><![CDATA[Facebook s&#8217;est, semble-t-il, fourvoyé avec l&#8217;inclusion de Facebook Connect sur des sites tiers. Facebook Connect la solution de SSO (Single Sign On, centralisation d&#8217;authentification) fournie par le géant Facebook. Vous connectez votre compte Facebook avec la communauté d&#8217;un site, d&#8217;un blog et pouvez commenter les billets et interagir avec les autres membres.
On se réfèrera au [...]]]></description>
			<content:encoded><![CDATA[<p><img class="mt-image-left" style="margin: 0pt 20px 20px 0pt;float: left" src="http://www.groupereflect.net/blog/w3c-facebook-thumb-125x125.png" alt="w3c-facebook.png" width="125" height="125" />Facebook s&#8217;est, semble-t-il, fourvoyé avec l&#8217;inclusion de <em>Facebook Connect</em> sur des sites tiers. <em>Facebook Connect</em> la solution de SSO (Single Sign On, centralisation d&#8217;authentification) fournie par le géant Facebook. Vous connectez votre compte Facebook avec la communauté d&#8217;un site, d&#8217;un blog et pouvez commenter les billets et interagir avec les autres membres.<br />
On se réfèrera au <a href="http://wiki.developers.facebook.com/index.php/Trying_Out_Facebook_Connect"><em>Facebook developer wiki</em></a> pour implémenter cette fonctionnalité. Cependant, le code généré et proposé sur le <a href="http://wiki.developers.facebook.com/index.php/Trying_Out_Facebook_Connect"><em>Facebook developer wiki</em></a> ne passe pas la validation W3C et ce, avec n&#8217;importe <a href="http://www.w3.org/QA/2002/04/valid-dtd-list.html">quel doctype XHTML</a>, le problème semble encore assez problématique avec le <a href="http://dev.w3.org/html5/spec/"><em>draft</em> HTML5</a>. En cause, l&#8217;appel à de multiples espaces de nom :<br />
<code>&lt;html xmlns="http://www.w3.org/1999/xhtml" xmlns:fb="http://www.facebook.com/2008/fbml"&gt;</code></p>
<p>Théoriquement, le tag &lt;html&gt; fait appel à 2 espaces de nommage (xmlns et xmlns:fb) et XHTML est sensé l&#8217;autoriser (au moins dans sa version 1.1). Mais <em>que nenni</em>, cette implémentation n&#8217;est que sur le papier&#8230; dommage !</p>
<p>Je n&#8217;ai visiblement <a href="http://stackoverflow.com/questions/599334/facebook-connect-wont-validate">pas été le seul</a> à constater cet état de fait et après une petite discussion sur Twitter (merci à <a title="Chez Jérémie" href="http://jeremie.patonnier.net/">Jérémie Patonnier</a> pour ses avis pertinents), il s&#8217;avère que l&#8217;insertion de <em>Facebook Connect</em><br />
à un site, fait perdre la validation W3C sans que l&#8217;on y puisse grand<br />
chose. La seule solution que j&#8217;ai pu entrevoir, est de redéfinir la DTD<br />
du doctype comme mentionné dans cet article d&#8217;A List Apart : <a href="http://www.alistapart.com/articles/customdtd/">Validating a custom DTD</a>.<br />
Cependant, cette technique ne sera pas intégralement reconnue par le<br />
validateur du W3C car s&#8217;appuyant sur un doctype non homologué&#8230; le<br />
serpent se mord la queue !</p>
<p>Même après de vaines recherches, le forum de <em>Facebook developer</em> reste désespérément sans réponse pour cette problématique.</p>
<p>C&#8217;est tout de même étrange que l&#8217;un des sites au plus fort trafic mondial (le 3ème selon <a href="http://www.alexa.com/topsites">Alexia</a>, 2ème en France) oblige à faire une telle gymnastique pour avoir un code propre mais non valide.</p>
<p>Facebook semble plus se soucier de ce qui fonctionne plutôt que de ce qui est valide. J&#8217;en arrive à une réflexion : doit-on se soucier de la validité du code de nos applications ? Je dis &laquo;&nbsp;oui !&nbsp;&raquo;. Cette affaire n&#8217;est pas sans rappeler quelques tentatives de flibustiers pour faire passer leurs propres implémentations de fonctionnalités navigateurs pour des standards de fait&#8230; suivez mon regard.</p>
<p>Chez groupeReflect, nous nous appuyons sur les standards et les bonnes pratiques afin de délivrer un produit de qualité, avec une forte accessibilité, utilisable sur une grande majorité de navigateurs et surtout, dont le code sera pérenne même si les navigateurs évoluent : c&#8217;est notre engagement !</p>
<p>Malheureusement, avec Facebook Connect, nous nous heurtons à un obstacle qui nous oblige à revoir la qualité de nos livrables <span style="text-decoration: line-through">sans pouvoir maudire</span> sans pouvoir mots  dire.</p>
<p>Avez-vous déjà rencontrer ce problème ? Pensez-vous que la gestion des standards ne sera plus l&#8217;apanage des fabricants de navigateurs mais également celui des sites à fort trafic qui pourrons tenter de faire avancer les choses plus rapidement ou d&#8217;imposer leur vision des standards ?</p>
]]></content:encoded>
			<wfw:commentRss>http://www.groupereflect.net/blog/archives/2009/08/facebook_et_les.html?parole_d_expert/feed</wfw:commentRss>
		<slash:comments>4</slash:comments>
		</item>
		<item>
		<title>La dinde au Whisky</title>
		<link>http://www.groupereflect.net/blog/archives/2005/02/la_dinde_au_whi.html?parole_d_expert</link>
		<comments>http://www.groupereflect.net/blog/archives/2005/02/la_dinde_au_whi.html?parole_d_expert#comments</comments>
		<pubDate>Mon, 28 Feb 2005 07:53:50 +0000</pubDate>
		<dc:creator>David Lafon</dc:creator>
				<category><![CDATA[Humour]]></category>

		<guid isPermaLink="false">http://wordpress.groupereflect.net/?p=77</guid>
		<description><![CDATA[Voici une blague gastronomo-éthylico-gallinacéenne &#8230; Certes, elle n&#8217;est pas toute jeune mais il semblerait que certains ne la connaissent pas encore ! Alors imaginez-vous dans la peau d&#8217;un chef cuistot, en préparation pour les &#171;&#160;Bocuses d&#8217;or&#160;&#187;, lors de la dernière ligne droite, la tension est à son comble, les nerfs à fleurs de peau, le [...]]]></description>
			<content:encoded><![CDATA[<p>Voici une blague gastronomo-éthylico-gallinacéenne &#8230; Certes, elle n&#8217;est pas toute jeune mais il semblerait que certains ne la connaissent pas encore ! Alors imaginez-vous dans la peau d&#8217;un chef cuistot, en préparation pour les &laquo;&nbsp;Bocuses d&#8217;or&nbsp;&raquo;, lors de la dernière ligne droite, la tension est à son comble, les nerfs à fleurs de peau, le stress fait monter l&#8217;adrénaline et&#8230;<br />
A vos marques, prêt, cuisinez !</p>
<p><span id="more-77"></span><br />
Recette de la Dinde au Whisky<br />
Ingrédients:<br />
Acheter une dinde d&#8217;environ 5 kgs et une bouteille de Whisky<br />
Prévoir du sel, du poivre, de l&#8217;huile et des bardes de lard.<br />
Durée de la préparation:<br />
Une bonne journée.<br />
Recette:<br />
- Barder la dinde de lard, la saler; la poivrer et ajouter un filet d&#8217;huile d&#8217;olive.<br />
- Préchauffer le four therm. 7 pendant 10 minutes.<br />
- Se verser un verre de whisky. Le boire.<br />
- Mettre la dinde au four dans un plat à cuisson.<br />
- Se verser 2 verres de whisky et les boire.<br />
- Après une demi-heure, fourrer l&#8217;ouvrir et surbeiller le buisson de la pinde.<br />
- Brendre la vouteille de buiscuit et s&#8217;enfiler une bonne rasade.<br />
- Après une demi-heure, dituber jusqu&#8217;au bour.<br />
- Ouvrir la borte, reburner, revourner, enfin bref, mettre la guinde dans l&#8217;autre sens.<br />
- S&#8217;asseoir une un butain de chaise et reverdir 5 ou 6 verres de whisky.<br />
- Buire&#8230;, non luire.. ., non cuire la bringue bandant 4 heures.<br />
- Et hop, 5 berres de plus.<br />
- R&#8217;tirer le four de la dinde.<br />
- Se rebercer une bonne goulée de whisky.<br />
- Ramasser la dinde (l&#8217;est tombée par terre). L&#8217;ettuyer et la voutre sur un blat&#8230; sur un clat&#8230;   sur une assiette.<br />
- Se béter la fibure à cause du gras sur le barrelage de la cuisine.<br />
- Ne pas essayer de se relever.<br />
- Décider qu&#8217;on est bien par terre et binir la mouteille de rhisky<br />
- Plus tard, ramber jusqu&#8217;au lit; dorbir ze gui reste de la nuit.<br />
- Le lendemain matin, prendre un Alka Seltzer, manger la dinde froide avec de la mayonnaise et nettoyer le bordel que vous avez mis dans la cuisine.<br />
Bonne chance&#8230; <img src='http://www.groupereflect.net/blog/wp-includes/images/smilies/icon_smile.gif' alt=':-)' class='wp-smiley' /> </p>
]]></content:encoded>
			<wfw:commentRss>http://www.groupereflect.net/blog/archives/2005/02/la_dinde_au_whi.html?parole_d_expert/feed</wfw:commentRss>
		<slash:comments>2</slash:comments>
		</item>
		<item>
		<title>Bill, tu dors ?</title>
		<link>http://www.groupereflect.net/blog/archives/2005/01/bill_tu_dors.html?parole_d_expert</link>
		<comments>http://www.groupereflect.net/blog/archives/2005/01/bill_tu_dors.html?parole_d_expert#comments</comments>
		<pubDate>Tue, 11 Jan 2005 11:40:01 +0000</pubDate>
		<dc:creator>David Lafon</dc:creator>
				<category><![CDATA[Techno]]></category>

		<guid isPermaLink="false">http://wordpress.groupereflect.net/?p=13</guid>
		<description><![CDATA[En ce début d&#8217;année, la firme de Redmond (Jim Allchin, responsable de la division Windows dans le Seattle Times) a fait une déclaration concernant notamment la sortie d&#8217;Internet Explorer&#8230; ahemmm, plutôt le retard que prenait l&#8217;évolution d&#8217;Internet Explorer. En effet, la prochaine version d&#8217;IE (6.5, 7.0 nul ne le sait !) ne devrait voir le [...]]]></description>
			<content:encoded><![CDATA[<p><img alt="bill-gates.jpg" src="http://www.groupereflect.net/blog/archives/bill-gates.jpg" width="159" height="125" hspace="5" align="left" />En ce début d&#8217;année, la firme de Redmond (Jim Allchin, responsable de la division Windows dans le <a href="http://seattletimes.nwsource.com/html/microsoft/2002124782_ie20.html" target="_blank">Seattle Times</a>) a fait une déclaration concernant notamment la sortie d&#8217;Internet Explorer&#8230; ahemmm, plutôt le retard que prenait l&#8217;évolution d&#8217;Internet Explorer. En effet, la prochaine version d&#8217;IE (6.5, 7.0 nul ne le sait !) ne devrait voir le jour qu&#8217;en 2006 avec Longhorn.<br />
Que penser de ce retard, volonté de sortir enfin un navigateur digne de ce nom, réel retard sur la réalisation de Longhorn et de ces logiciels ou simplement je-m&#8217;en-foutisme total ?<br />
Aux vues de l&#8217;interview que Bill Gates a donné durant le Consumer Electronics Show de Las Vegas (<a href="http://fr.news.yahoo.com/050107/7/47lvv.html" target="_blank">ici</a> et <a href="http://news.com.com/Gates+taking+a+seat+in+your+den/2008-1041_3-5514121.html?tag=nefd.ac" target="_blank">là</a> en anglais), on est en droit de se poser des questions quand à la lucidité des dirigeants de Microsoft. La démagogie dont fait preuve BG dans cette interview est à la limite de l&#8217;insulte à l&#8217;intelligence (voir le très bon papier de Michel Dumais de <a href="http://www.ledevoir.com/2005/01/10/72213.html" target="_blank">ledevoir.com</a>). Il est prêt pour la politique !<br />
En attendant, la croissance incroyable du nombre d&#8217;utilisateurs de Firefox ne semble pas l&#8217;inquiéter, tout comme les failles de sécurité d&#8217;IE qui ne sont toujours pas patchées après plusieurs semaines&#8230; brrr, ça fait froid dans le dos.<br />
Non, je ne fais pas de &laquo;&nbsp;l&#8217;anti-microsoftisme primaire&nbsp;&raquo;, mais force est de reconnaître que leur stratégie de séduction du public est de plus en plus éloignée de ses aspirations (&#8230; du public &#8230; faut suivre ;-p ). Plus je me pose des questions sur l&#8217;avenir du logiciel et plus je me dis que cet avenir passe par le libre et non par une société qui gère simplement et aveuglement son monopole. Faut-il se retourner vers Firefox, Thunderbird, Mandrake et consorts qui maintenant offrent des fonctionnalités et une ergonomie qui égalent voire dépassent Windows ? Je dirais oui &#8230; mais vous ?</p>
]]></content:encoded>
			<wfw:commentRss>http://www.groupereflect.net/blog/archives/2005/01/bill_tu_dors.html?parole_d_expert/feed</wfw:commentRss>
		<slash:comments>0</slash:comments>
		</item>
	</channel>
</rss>
