eValid™ -- Automated Web Quality Solution Browser-Based, Client-Side, Functional Testing & Validation, Load & Performance Tuning, Page Timing, Website Analysis, and Rich Internet Application Monitoring. © Copyright 2000-2011 by Software Research, Inc.

Summary
Tests of a web server done from the client side -- i.e. from inside a browser -- simulate user interactions by stimulating the server to produce and deliver various pages based on requests made by the client (the user). eValid's in-the-browser spider examine every page viewable and analyze the delivered contents. The results are analyses that are extremely accurate and 100% indicative of what users experience.

Main Advantages of Logical View
The main advantages of working from the "logical view" of a website using eValid's unique browser-based approach to website analysis and quality assurance include:

• Accurate User Perspective. The analysis is done entirely from the point of view of a user, based on the information that is drawn from the web server into the eValid browser by requests made by eValid's automatic search spider.

• Accurate Timing. Timing data is 100% representative of what the user sees because the measurements are done 100% from the client perspective. Download times includes the time it takes to fully download and render the page.

• Accurate Sizes. Size calculations for pages, constituent elements, images, and other files, are always based on the information actually delivered by the server to the evalid browser.

• Accurate Dynamic Content. Only content actually generated by the server is analyzed at the receiving end. evalid's browser based approach is immune to all server implementation details, and only analyzes the material generate in response to a browser request.

• Actual Link Dependence Structure. Some links may not be explicit in file-based scans. Only actual active links are used by eValid. Parent and child relationships between pages, and between pages and page components, are 100% accurate.

• Handles Secure Protocols. Special protcols, including HTTP/S, require no special modification or support.

• Handles Browser Plugins. The client-side approach handles browser plugins and all other types of client-side appliances.

• Accurate Metrics. Metric data collected for pages reflects the information known to the eValid browser, within the Document Object Model (DOM). (Metric data other than basic size and download time is not usually collected for page components such as *.gifs and *.jpgs.) As a consequence, page complexity details pertain specifically to what is seen by the user, and is not based on server-internal complexity.

• Handles Login Sites. If a website requires a login, eValid's functional testing component can play back the login sequence and issue the site analysis start command after successfully logging in.

• Special Page Scan Logic. Because the entire delivered page is available for analysis, the completely HTML, visible text, META tags, or TITLE tag can be scanned for presence or absence of strings or regular expressions.

• No Special Access Rights Required. Because the analysis is done with information accessible by the public, no special access rights are required to read server-side files.

The Logical View
What happens when, in a web page, you click on a hypertext link to request that the server deliver you (the browser) a new page. The server generates the page and sends it to the browser. In turn, the browser may request additional pages based on what it was just sent. Modern browsers -- eValid included -- launch multiple threads to help get the complete page to you as fast as possible. For example, pictures (images) or cascading style sheets or JavaScript files are all requested automatically by the browser if the page to which you just navigated references them. (There are many ways a base page can make requests for secondary pages.)

Modern web servers generally compose the page that you see by assembling the page from many components, fragments, parts, and pieces. What you see are pages; you don't see the components that the server used to create them. (In older, so-called flat-file sites, every page does correspond to a file on the server.)

As the picture below shows, web page dependencies and structures are NOT the same as files. WebSite analyzers that operate on the server side use files which may or may not be actual visible, browsable entities. Note that none of the entities shown is specifically a "file" but instead are all logical units as seen from, and processed by, the browser.

Client Site (Browser Based) View Of Web Pages

The diagram illustrates the logical structure of a website and is the basis for the above discussion.

• Page Link. Page 1 (URL1.html) has a link to Page 2 implemented in a <A HREF=.. > tag. Page 2 is a child of Page 1.
• Parent Page Links. There are possibly many other links from other pages [parents] to Page 1 or Page 2. On websites pages can refer to themselves (making them their own parents), or they may refer to pages that refer to them (circular dependence).
• Child Page Links. There are possibly many other links to other pages [children] to Page 1 or Page 2. On websites pages can refer to themselves (making them their own children), or they may refer to pages that refer to them (circular dependence).
• Component Links. Both Page2 and Page2 have of connections -- which are also called -- to constituent files via a variety of HTML methods such as: LINK..., SCRIPT..., and IMG... among many others. In most cases components don't have links to other parts of the website.