Dynamic content, how many variables? How big can the variable be?

Is s=5&c=abc better than sitenumber=5&company=abc?

[Clark]

Has anyone done detailed analysis of how google deals with dynamic content based on changing variables?

We all know that sessionids suck.

But is a smaller variable name preferable to a long one, iow, "s=" is better than "sitenumber="?

Is one variable better than two?
i.e.
"sitenumber=5"
is better than
"action=sitenumber&sitenumber=5"
?

[mikemcs]

Good question I have also read in a post date 2 years ago in a do / do not thread that you should not have a? in your URL but I assume that has all changed.

[killroy]

It has for modern bots.

But still, if your using them the wrong way you will get in trouble sooner or later.? QUERY strings are jsut taht, QUERIES. if you use them to simply identify pages to pick you are ABUSING and should look into either server side URL modification (mod_rewrite) or at least use PATH_INFO instead like this:

domain.tld/scriptname/var1-val1/var2-val2/var3-val3 and so on...

Use the tools how they were intended and you are less likely to be misunderstood by others (i.e. Google)

SN

[Clark]

I didn't follow you when you said to PICK YOU, what did you mean?

I think if you have real content and want to make the url more friendly to google (like by removing sessionid info from the url), you are neither hurting google, nor your web page visitors, nor is it abuse. But I'm not really sure what you mean.

[killroy]

Was suposed to be "...pages to pick, you will..." sorry, my spelling went down the drain that post. Was late ;)

What I mean is, that query strigns were created for a specific purpose. And it wasn't to pick on page out of a set of pages that a script can display (like a product catalog). Rather it's to actually query a database, like for a search result, in a way that can throw a different result each time.

For selecting pages (like from a product catalog) it's better to use PATH_INFO or mod_rewrite. Please do a site search for those, to see the millions of threads on the topic. Also search for "Cool URIs don't change" on Google, it's a very enlightening article by the W3C.

SN

[sullen]

Using querystrings to show different content is not abusing anything. But for Google to crawl the pages its best to keep the info in the querystring to a minimum. So Clark, your first option is better.

Also Google doesn't always crawl all dynamic content (any page with a?) - it depends on the size of your site an your PR (apparantly). But I have a PR5 site with 20 pages like this and Google has crawled all of them just fine.

[killroy]

well its abuse in the same way using a car to flatten a can is abuse. You can do it, it won't cause any harm and it's not illegal... doesn't mean it's a smart thing to do and the right tool for the job.

That's what I mean by "abuse". 95% of modern dynamic pages use QUERY_STRING for the purpose that PATH_INFO was invented for. As the W3C recommendations say (and may I remind you that they're the ones who come up with the standards), QUERY_STRING is for ephemerial information that is NOT well defined by the URL + QUERY_STRING. A modern dynamic website delivers content pages (plus perhaps some minute dynamic info). The dynamic part comes really just form the fact that the content is loaded and combined on the fly, NOT that the content is unique to this request.

I'm not preaching rules and regulations, I'm merely promoting best practices, in the hope that one day the internet will be what it can be.

SN

[Clark]

Regarding killroy's comments, I agree that the web can be much more than it is. The promise of the semantic web is great, let's hope to see more of it implemented. Much of it is in the hands of google right now. If they choose to implement a standard, it will become useful. Whatever they support becomes a defacto standard. And there are many fine unsupported ones they can look at implementing.

I've used both Mod Rewrite (which I hate, because it is often unpredictable and doesn't work the way it is supposed to) as well as the querystring with great success on Google. Lots of dynamic content spidered.

But I don't like the idea of presenting different content for google and for the web visitor (unless it uses xml and the like in the manner it was built for...but google does not do that today). If the web visitor consistently visits domain.com?page=1, because page=1 is something generated by a database, it would be nice for google to follow the same link, along with the requisite PR, Anchor Link text values and other google ranking data that link deserves.

By using Mod Rewrite to duplicate the page, I'm not sure it really makes the internet a better place... Anyways, that's all semantics...I don't have time for Utopia anyways, unfortunately, it's about practicality.

The purpose of the thread was to find out what works because those longer queries aren't working as well for me so I will be try reducing the length of the url. Thanks sullen!

[Jesse_Smith]

s= better not be numbers if you want it listed, it will give you the Google Death Penalty. Google will think it's a session ID. I got rid of s=######## and bang, my vBulletin board started geting everything indexed for the first time in almost two years.

[plasma]

Same here.
Since I started to turn the query-string into a looks-like-static url google loves my sites :)

Apart from that, I also recommend using hashed urls to prevent ppl from fiddling around with the parameters.

But take care that pages with different hashed-urls dont reference the same content, otherwise you will get duplicate content.

[killroy]

Clark, I'm at a total loss regarding your coments about duplicating content. NOTHING needs to be duplicating. I'm simply wondering what is wrong with
domain.tld/content1
instead of
domain.tld/script?page=1

which do you like better? is the former hard to do? NO! Do you need to use mod_rewrite? hardly. In fact if you hate it so much, simply create ONE rule that pumps ALL urls into ONE script like>
domain.tld/content1 --> domain.tld/script/content1
I used that method for many HUGE sites, where the one script then figures yout the real URL. no mod_rewrite rules for all pages, and you can do with your URLs whatEVER you want. Even dynamic .js and .css files... The posibilities are endless...

Please donT' think me a nut case, I wouldn'T recommend this stuff if it wasn'T any good ;)

SN

[Clark]

Hey Killroy,

Sorry, if it sounded like I was being antagonistic to you, that wasn't how I meant it. I respect you, you seem to be pretty knowledgeable...

I guess I misunderstood PATH_INFO...are you saying that Path info does not use Mod ReWrite. I dunno, I keep learning more and always feel like I know nothing...

Hashed urls? Another one I don't get..