Forum > Content Management Systems > Topic

Content Management Systems Forum

How can we review the analyzer?

Marc Laporte on 15 October 2011, 14 years ago

Hi!

"We use a combination of regular expressions and DOM traversal to identify technologies. A lot of research was necessary to build the analyzer, and we keep improving it all the time. We want it to be the best possible website analyzer."

Source: https://w3techs.com/faq

How can we review this and make sure the stats are OK for a specific CMS?

Thanks!

Sam Soltano (site administrator) on 15 October 2011, 14 years ago

Hi Marc,

You can verify the results by checking what we report on specific sites. You can do this by entering any URL here: w3techs.com/sites.

Any reports of false-positive or false-negative results are highly welcome, and so are any specific hints on what should be considered in the analyzer.

Marc Laporte on 25 October 2011, 14 years ago

Hi Sam

Thanks for this.

Here is one: http://droit-inc.com

It reports as PHP, which is accurate, but omits the CMS. You can check the source for:  

In some older versions, you could also see these variants: 

While you are at it, can you update w3techs.com to use "Tiki WIki CMS Groupware" instead of "TikiWiki"? It would be nice to keep the old URLS to redirect to the new ones.

Best regards,

M ;-)

Sam Soltano (site administrator) on 25 October 2011, 14 years ago

Hi Marc,

Thank you very much for updating us on the change of the tool name and the generator tag. We have changed our data bases accordingly. Due to caching it will go online only in a few hours. Also, allow our crawler a few weeks to revisit all the sites whose Tiki Wiki usage may have remained undetected.

Thank you for your support, all the best

Sam

Marc Laporte on 25 October 2011, 14 years ago

Hi again!

Sorry for messy previous post. The WYSIWYG editor was tricked by my HTML :-)

So it's fine to keep tikiwiki as the system name for the URLs:

https://w3techs.com/technologies/details/cm-tikiwiki

But with full project name.

Best regards,

M ;-)

Sam Soltano (site administrator) on 25 October 2011, 14 years ago

Sure. That's what we did. We never change our URLs.

Marc Laporte on 25 October 2011, 14 years ago

Ok, super!

I see "I" instead of "i" in Tiki WIki CMS Groupware

Thanks!

Marc Laporte on 25 October 2011, 14 years ago

Here is a false positive for WordPress: https://w3techs.com/sites/info/designpedia.net

Best regards

Sam Soltano (site administrator) on 25 October 2011, 14 years ago

That's a good find. They use (parts of) a WordPress theme on Tiki Wiki. Is that a common way of designing a Tiki Wiki site?

Thanks for pointing out the misspelling.

Marc Laporte on 25 October 2011, 14 years ago

I wouldn't say it's common but this happens when people like the look & feel of their site but they want something with more features. (We have a data importer from WordPress)

I'll brag a little here :-)

http://tiki.org/FOSS+Web+Application+with+the+most+built-in+features

I found another:

https://w3techs.com/sites/info/wikiwackyworld.com

Checking for lib/tiki-js.js will detect it unless javascript is minified

Thanks!

Marc Laporte on 25 October 2011, 14 years ago

Hi!

Here is a somewhat customized Tiki:

http://www.austms.org.au/tiki-read_article.php?articleId=137

One way to detect it is the presence of files/link to files which have start with tiki- and end with .php 

There are hundreds of files with this pattern and no other CMS would use this.

http://tikiwiki.svn.sourceforge.net/viewvc/tikiwiki/trunk/

Thanks!

M ;-)

Sam Soltano (site administrator) on 26 October 2011, 14 years ago

That's all very helpful, Marc. We will have a close look at all of these cases.

Marc Laporte on 26 October 2011, 14 years ago

Here is another: 

https://w3techs.com/sites/info/collaborative-translation-patterns.com

it reports II6/ASP.net instead of Tiki

Thanks!

Sam Soltano (site administrator) on 28 October 2011, 14 years ago

Hi again,

We have included your valuable remarks into our analyzer, which should somewhat improve the TikiWiki recognition.

Just a side note: two of your examples have other issues. One uses a JavaScript trigger to redirect to another page. Although we often recognize JavaScript-based redirects, we don't execute JavaScript triggers, so that was beyond our limits.

Another one uses Frames. That by itself is not a problem, as we also step into frames, but in that case the framed page was from a completely different domain, and that's where we draw the line. We don't consider that page to be part of the site anymore.

Now, one can discuss whether these policies are appropriate or not, but what is important for us is, that these decisions are independent from any CMS. If we don't report such sites as TikiWiki, then the same will happen to other CMSs, so that our surveys are not distorted.

You need to be logged in to reply.