TIOBE D ranking

[Thomas Kuehne]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Kyle Furlong schrieb am 2006-03-18:
> Thomas Kuehne wrote:
>> 
>> BCS schrieb am 2006-03-17:
>>> James Pelcis wrote:
>>>> As of March 2006, D is in 21st place!  In comparison, this is just short 
>>>> of VB.net, despite the major difference in advertising power. Go D!!!
>>>>
>>>> http://www.tiobe.com/index.htm?tiobe_index
>>>   WOW! Almost on the first page, and there are a lot of big names further down:
>>>
>>> Ruby, Fortran, Smalltalk, OCaml, Objective-C 	
>> 
>> TIOBE is pretty anglophil.
>> 
>> e.g.: their search critera don't count the many Japanese Ruby pages
>> 
>> In addition "D" isn't a good keyword:
>> 1) mark "D"
>> 2) part "D"
>> ...
>> 
>
> They don't just search for D silly, they search for "d programming"

I didn't say their data is flawed - just be careful when interpreting
the data(as with every other statistic).

1) "programming ... -tv"
* due to "programming" only English pages will be found
* only pages that aren't served from a .TV domain will be found

2) "programming <language> ..."
* false hits vary by "<language>"
+ "programming ruby": no significant amount of false positives
+ "programming Java": includes all software shops in Java(the location)
+ "programming C/D": lot's of false positives(truck loads of programming
documentation, talking about CS grades, ...)

3) Well documented OpenSource and academic projects have a much larger
online presents that links "programming" and "<language>" than
commercial and in-house software.

I'll stop here, but there are several other points that have to be
considered while analyzing the TIOBE data.

Thomas


-----BEGIN PGP SIGNATURE-----

iD8DBQFEHDBY3w+/yD4P9tIRAkJuAKCt1Iqnlul91CsRjkR3rXrnJysSQACeIoq7
+l0LHRR9l0ysEzOKtJYhzcA=
=WBPQ
-----END PGP SIGNATURE-----