Comments on: Page Speed 1.6 Beta – new rules, native library

By: Vincent Flanders

Vincent Flanders — Thu, 04 Feb 2010 00:56:43 +0000

The Minify HTML seems to be screwed up and mangling the pages (although they display correctly). Validate http://www.webpagesthatsuck.com/articles.html — which has 3 errors caused by Google — against the “minified” version http://www.webpagesthatsuck.com/articles-screwed-up.html — which has 577 errors.

By: Steve Souders

Steve Souders — Tue, 02 Feb 2010 23:12:51 +0000

@kangax: Documentation for these new rules should be there soon. (Our tech writer was on vacation!)

By: kangax

kangax — Tue, 02 Feb 2010 22:49:56 +0000

@Billy Hoffman

Actually, removing quotes around attributes or dropping optional tags does not always affect document validity. If we are talking about HTML 4.01 (which most of the documents on the web are _served_ as, even with misleading XHTML doctype stamped on top), there’s absolutely nothing invalid about, say, “P” or “LI” elements with omitted closing tags — HTML 4.01 doctype allows it (http://www.w3.org/TR/REC-html40/sgml/dtd.html). You might have noticed that even in “Agressive optimizations” section of my post (http://perfectionkills.com/optimizing-html/#agressive_optimizations), I mentioned that none of the transformations invalidate a document.

@Steve

Clicking on “Minify HTML” in PageSpeed brings to a page that has no such section. PageSpeed also always shows “ok” icon, even when on non-minified pages. Am I missing something? What’s the exact criteria for this rule to start alerting about performance problem?

By: Sergey Chernyshev

Sergey Chernyshev — Tue, 02 Feb 2010 22:25:28 +0000

WOW – I missed native library part in 1.15!

I’d love to have a daemon that runs rules and then sends them as a beacon to ShowSlow and all that without a browser – that’d be just perfect!

By: Steve Souders

Steve Souders — Tue, 02 Feb 2010 17:25:42 +0000

@Ofer: There's room for minifying HTML and still passing validation. I think it would be good to have an option in the minifier to do this, or to be more aggressive if the developer wants to. I often use the W3C validator. Right now, you should run both Page Speed and YSlow. There are about 25 rules that overlap, but each tool has 5-10 rules that are unique, and some of these rules are critical. That's why I'm excited about the native library. We could move in a direction where there's a common set of rules that is shared by all tools, so developers don't get varied results depending on their choice of OS, browser, and tool.

By: Billy Hoffman

Billy Hoffman — Tue, 02 Feb 2010 14:55:35 +0000

sorry, that posted prematurely! How embarrassing! :-)

I meant to say that basic removing of whitespace and removing HTML comments that are not conditional comments should not affect validation at all.

By: Billy Hoffman

Billy Hoffman — Tue, 02 Feb 2010 14:54:06 +0000

Ofer,

1- It depends. If by “minify” are simply removing HTML and non-conditional HTML comments than it will not effect validaiton. Some commerical/shareware HTML minifiers will do things like removing closing tags (like ), removing quotes from attributes, things like that. Those you should stay away from because will break validation.

Check out this article for more safe HTML minification tips: http://perfectionkills.com/optimizing-html/

2-HTML Tidy should do basic minification no problem. Just be careful when running it on PHP or ASP.NET files instead of pure HTML files. It might strip out some of the server-side application code!

By: Ofer Wald

Ofer Wald — Tue, 02 Feb 2010 08:49:44 +0000

I have a question here, wrt your stance on the html minify feature.
1. won’t minified pages fail validation?
2. any programs that actually do that?

which ruleset is better, yahoo’s or google’s?