Web Directions Year round learning for product, design and engineering professionals

• Server Hardware
• Server OS and Web Servers
• Databases
• Back end Programming languages and Frameworks

In our final section, we asked about server side languages, frameworks and technologies respondents used. It included questions about server hardware, server operating systems, web servers, database systems, and backend programming languages and frameworks.

Server Hardware

When asked “What type of hardware do your sites run on?”, 35% responded “Dedicated servers managed by you or your company”, 30% shared servers, 15% “Dedicated co-located servers”, while 5% said “don’t know”.

What type of hardware do your sites run on?
Answer	Count	Percentage
Dedicated servers managed by you or your company	433	35.09%
Dedicated co-located servers	189	15.32%
Shared servers	368	29.82%
Don’t know	62	5.02%
Other	40	3.24%
No answer	17	1.38%
Non completed	125	10.13%

Server OS and Web Servers

Linux and Unix dominate the server side operating systems used, with 59% and 17% respectively, while Windows weighs in at 28%. Perhaps surprisingly, Mac OS X is used as the server side operating system by nearly 6% of respondents.

What server software is used to serve your sites?
Answer	Count	Percentage
Apache	849	68.80%
IIS	280	22.69%
GWS	5	0.41%
lighthttpd	68	5.51%
Don’t know	115	9.32%
Other	81	6.56%

As to be expected, Apache and IIS dominate the web server market, with Apache at 69% and IIS at 23%. As with the operating system figures, this is a much greater discrepancy than between these servers and operating system than is generally reported by surveys like Netcraft’s (which puts Apache at around 48% and IIS at 33%). Lighthttpd came in at 5.5%, while Google’s GWS is very lightly used by comparison with other surveys.

It needs to be kept in mind that server software and operating system surveys like Netcraft’s typically measure market share in terms of sites served, where this survey is essentially per developer. As such, any comparison between market share from those studies and this are probably of little meaningful value.

Which operating system(s) do your servers run?
Answer	Count	Percentage
Linux	725	58.75%
Unix	206	16.69%
Windows	343	27.80%
Mac OS X	69	5.59%
Don’t know	113	9.16%
Other	26	2.11%

Databases

When asked “What database systems do you use?”, only 3.4% of respondents replied “none” – further emphasizing the increasing move away from static page based sites to dynamic sites.

Among the database systems nominated by respondents, MySQL dominates with over 70% of developers using it. Microsoft’s Microsoft SQL Server has a 22% market share among respondents, PostgreSQL 10%, and Oracle 9%. Other databases nominated by respondents include SQLite (around 1.5%), with a handful using CouchDB, Filemaker, and Access.

What database systems do you use?
Answer	Count	Percentage
none	42	3.40%
MySQL	869	70.42%
Microsoft SQL Server	271	21.96%
Oracle	113	9.16%
PostgreSQL	128	10.37%
Other	80	6.48%

Back end Programming languages and Frameworks

When asked “Which programming languages do you use?”, only 3.5% of respondents replied “none”, once again emphasizing the increasingly dynamic nature of web sites we’ve see from the high percentage of developers who use JavaScript, and database systems.

Of the languages nominated by respondents, PHP dominates, at 63%. JavaScript comes in at a surprisingly high 55%. ASP.NET is used by17% of respondents, Python 15%, Ruby, despite it’s current high profile, 14%, ASP and Java 12%, and Coldfusion at 7%.

Of “other” languages, a handful of respondents use ActionScript, with few other languages being used by more than one or two respondents. As with all our other questions, the full details are available to download.

Which programming languages do you use?
Answer	Count	Percentage
ASP	144	11.67%
ASP.NET	208	16.86%
ColdFusion	88	7.13%
Java	152	12.32%
JavaScript	680	55.11%
Perl	103	8.35%
PHP	778	63.05%
Python	187	15.15%
Ruby	178	14.42%
None	43	3.48%
Other	43	3.48%

The use of these languages increasingly goes hand in hand with a framework. These frameworks are language specific – and for some languages, such as PHP, there may be more than one widely used framework.

Interestingly, 29% of respondents answered that they use no framework for back end development. Compare this with only about 5% of respondents who use JavaScript but who don’t use frameworks.

Of those who do use frameworks, 13% use Ruby on Rails (despite Ruby being quite a way down the list of languages, this is the single most used framework. This may be due to the fact that most Ruby developers likely started using the language because of the Rails framework.)

Other commonly used frameworks include Django, the Python Framework (11%), the Zend Framework (the most commonly used PHP framework, at 6.4%, but used by only 10% of PHP developers in this survey), and CakePHP, a close second to Zend among PHP frameworks at 6%.

Other frameworks mentioned include the Java framework Struts (about 2%), the PHP framework Code Igniter, at about 1%, and Spring for Java at about 1%. A number of content management systems, such as WordPress, Joomla, and Drupal, which while not technically speaking frameworks, are increasngly being used in ways similar to them.

Interestingly, just under 20% of respondents said that they use internally developed frameworks.

The disparities between the use of libraries and frameworks for JavaScript and back end development are certainly interesting. Not just the high percentage of respondents who don’t use frameworks for back end development, but do for JavaScript development, but also the high proportion of those who use their own frameworks for back end development, and yet the near absence of internally developed frameworks for front end development.

Which back-end frameworks do you use for development?
Answer	Count	Percentage
None	367	29.74%
CakePHP	75	6.08%
CherryPy	3	0.24%
Django	133	10.78%
Ext GWT	4	0.32%
Ruby on Rails	164	13.29%
Struts	23	1.86%
Zend framework	79	6.40%
Developed internally	241	19.53%
Other	199	16.13%

Next

We’ll wrap up with some conclusions, and try to get a sense of the state of the web at the end of 2008.