81 replies to this topic

[jehochman]

Matt's posted a thread on his blog where he's asking us to vote for features we most want for Google Webmaster Tools. One of the proposed features would tell the webmaster which pages on their site don't validate.

Why would Google have that information? Are they validating pages just for the fun of it? I found this rather shocking because most of us think validation has nothing to do with rankings. Maybe it has something to do with indexing, or the method of indexing used.

In Providence last month Dan Crow told us that Google ability to index the Internet is limited by power and bandwidth, not money. He suggested that Google only allocates so much time and bandwidth to spider each site, so for large sites it's in the site owner's interest to do things that make the indexing process more efficient, because that may lead to more pages being indexed.

One way to make a site easier to index is to reduce code bloat. Another might be clearing coding errors so that the indexing program can quickly figure out the page, rather than having to make guesses when the code is all screwy. Guessing and compensating for errors takes time, processor power, and ultimately electricity. When indexing billions of pages, trivial things that eat up CPU cycles can create a significant burden.

[Michael Martinez]

I think validation is necessary or at least helpful with Accessible Search (a point I have neglected to address in past discussions about validation). It makes sense that Google would want to ask Webmasters if that is important to them.

[mcanerin]

I'm not sure that Google is validating already and just not reporting it. Many of the other options are related to things Google isn't doing now, as well, such as a tool to move from one domain to another. I was looking at it as an extension/variation of an earlier choice, "Score the crawlability or accessibility of pages".

Come to think of it, this is all assuming that he meant W3C validating, and not some sort of "Google" validating, which may or may not be more useful, depending on your viewpoint, and would not surprise me given Googles infamous "we can do it better than anyone else" approach to things.

Ian

[1dmf]

lol - non standards / invalid code might be coming to bite some people in the behind sooner than you think

Come to think of it, this is all assuming that he meant W3C validating, and not some sort of "Google" validating,

let's hope they don't go down the path MS did with standards making it up as they went along in IE6 , it drives us coders mad when you have to incorporate different code for different browsers!

Crickey if it became a choice between W3C standards and not getting indexed or Google standards and get indexed, all hell is gonna break loose in the 'standards compliant' community, and I for one won't be happy having just W3C'd all my sites - oh well , might keep me out of mischeif having to serve different pages for Google than the Joe public - oh hangon - isn't that 'black hat' SEO.

And to what standard does google's website validate to because it certainly isn't W3C - Stop it my sides are hurting

Edited by 1dmf, 23 July 2007 - 05:31 PM.

[Jill]

Why would Google have that information?

I wouldn't read anything into it, Jonathan, other than people have probably been asking Matt for that sort of thing because they read in so many places that valid code will help rankings. (Which we all know, it doesn't.)

[jehochman]

But Google has limited computing power and bandwidth. Why would they waste it to provide this feature? It doesn't make sense unless they're already doing the calculation for some other purpose. Perhaps this is just a red herring and they have no intention of providing the feature.

[mcanerin]

Or they intend to scrape the results from the W3C, using THEIR bandwidth, etc...

Ian

[maleman]

(Which we all know, it doesn't)

But... if G has limited computing power and bandwidth, it may use up whatever is allocated for a given site crawl on a few bloated pages and split.

I always try to keep my main pages short and sweet (300 to 400 words) and get them to validate to whatever doctype. Also I put intersite links to those pages where they can be found fast.

Edited by maleman, 23 July 2007 - 08:41 PM.

[projectphp]

But Google has limited computing power and bandwidth. Why would they waste it to provide this feature?

The answer is simple: they don't need to recrawl your site to validate it. They just need to run a parser over the downloaded HTML.

It doesn't make sense unless they're already doing the calculation for some other purpose.

The parser is likely open source open source, is wrtten in Perl, meaning it would be trivial to change it to read from a database / fielsystem, and spit out output into the Webmaster Central template, and already is pretty battle tested.

Google need never know the outcome, and might not even bother to record the output / result.

The time thing is pretty obvious, and pretty useful to know. One wonders why so many sites are so code heavy. Stripping out newlines, tabs and commetns always seemed a good idea to me.

[Jill]

Or they intend to scrape the results from the W3C, using THEIR bandwidth, etc...

Ian

Exactly.

Anyway, who says they're going to offer it? They're just 'askin what people think are important, no?

[1dmf]

It certainly would seem odd, re-inventing the wheel, when the acknowledged standards is W3C , who offer validating services already and seeing as they set/draft the standards, I think google simply need to stick to what they are good at.

It wouldn't hurt for there to be a button next to each page in the Webmaster Tools section where you could click validate and it just passes this to W3C in a new window, leaving the W3C to validate with their tool and bandwidth and google having a quick and easy access option in the WMT's.

Maybe they could even record the result next to each page, then Google wouldn't have to validate while crawling simply cross reference against their DB to see if a page is valid. Then do what ever accordingly, seem to me a more sensible approach.

[piskie]

Or just maybe they are already either doing it or some part way towards it (serious time consuming errors only)and as such want to know if it would be a popular addition to offer.

Code Bloat and Errors could be a factor if not now then sometime in the future. Head in Sand Merchants can ignore the "Possibility" at their own risk.

"They will never do that" is a brave assumption because after all Never is a long time.

[projectphp]

Seriously, why theorise without investigasting? The W3C validator is open source (I don't add links pointlessly), which means Google are free to use it (here is the licence), to validate pages they have already downloaded. Minimal bandwidth costs. No development costs (beyond changing a template and a simple DB call). Seems a useful tool to offer and a PR win for minimal output.

Occam's razor: simplest answer is often the best.

No conspiracy. No new knowledge. Nothing to learn here. Move along.

[1dmf]

I for one couldn't be doing with keeping open source code up-to-date everytime it changes, a form that submits to their already compiled and up-to-date website code seems far more logical.

But hey, like you say nothing to learn here.

Edited by 1dmf, 24 July 2007 - 08:41 AM.

[projectphp]

Huh? I have no idea what that means! I don't even know who "they" are: GOogle or W3C? I don't know either what " form that submits to their already compiled and up-to-date website code" means.

I think you are a bit confused as to what the code does, but then, not looking at the code creates that issue