Been using HTMLValidator for a while

Mostly too busy at the moment to do much about it. But it would be nice to convert a validator to run within Firefox. Looked at the W3C validator at one time, and it's written in C - not something that's easily translatable.

Problem I have with most of the validators is two fold. First, the external validators are hard to use in web applications (kind of a spinoff on Ben's suggestion about cookies, though I run a local web server off my box for development). You basically have to save off the page as an html file, put it in a publicly accessible location and then run the validation (I'll second Ben's suggestion of cut and paste HTML into a textarea for validation).

Second problem I have is that the validation doesn't usually look at the the DOM after the javascript cranks through. What I'd like is to verify the DOM at any particular point in time, not just at start up. I've got a bit of dynamic client side stuff happening that is invisible to all the validators that I've used.

Anyhow, Scrutinizer looks good. I assume you've written it in Smalltalk since it refers to Seaside?

that posts the page's source code to a url. Firefox also has "view rendered source" which, if you can hook that, would give you the post-javascript html.

I did write it in Smalltalk - its the only way to get this much done by yourself. It also makes it dead easy to install (one squeak vm, one image, one changes file) and it runs on just about anything. It is also possible to push changes from my dev image on my laptop to the deployed one without taking the app down.

I have many issues with the w3c validator. Installation is fiddly and you have to have all latest C libs on your system (its actually written in Perl with dependencies on a C SGML library which depends on the latest glibc....). The output of the thing breaks my browser when using it on my employer's home page. IOW, you see the big list of errors, click it to go to place in source code and occasionally the browser loses its mind and crashes. I couldn't reliably find errors with it. Additionally, it doesn't classify errors by severity - rather it lists them in the order in which it finds them and I think it complains too much about stuff that isn't that important - like escaping ampersands in urls in href attributes. Nobody bothers to do that and the browsers are more than OK with it.

Structural errors, OTOH, can result in totally broken pages and I really want to fix those.

Other stuff that nothing really catches that well is script syntax errors - we have deployed stuff with minor syntax errors that resulted in very visible error dialogs o some browsers. Something needs to check those.

I appreciate your feedback.

Welcome to IWETHEY!