survey results: reasons for choosing a database

[blu-5a1Jt6qxUNc at public.gmane.org (Tom Metro)]

On page 8 of this issue of "SD Times" magazine:
http://www.sdtimes.com/download/images/sdtimes179.pdf

is an article titled, "SQL Server Still No. 1 in Databases," which 
publishes the results of a survey of software development managers.

It includes a chart titled, "Which were the top factors that led to the 
decision to use the database in your current project?" It is interesting 
to note that all but one of the items cited among the top 10 reasons for 
choosing a database are non-technical:

45.9% Familiarity with database
21.3% High availability or reliability features
20.1% Lowest development costs
18.6% Lowest deployment costs
17.1% Covered under site license
15.3% Reputation of vendor
15.3% Required by specific applications
14.5% Required by customer/partner
14.5% Our legacy applications required it
12.3% Integration with app server

(I consider "High availability or reliability features" to be the one 
technical item above.)

So the top factor is "Familiarity with database," showing the impact 
simple popularity has on database choice. On top of that, the top 5 
includes "Lowest development costs," which I'd interpret as also 
including the cost of finding and hiring developers that know the 
product, and "Lowest deployment costs," which again indicates the value 
placed on using an already widely deployed product.

You have to dig deeper in the list to start finding technical reasons 
for choosing a database: #13 Had best database administration tools, #17 
Integration with IDE, #18 Had best development tools.

And really deep to start finding traditional core database technical 
abilities cited as the reason: #20 Lowest hardware requirements, #21 Had 
highest transaction performance, #23 Smaller databases lacked 
performance, #24 Smaller databases lacked key features, #31 Lowest 
memory footprint requirements. Each of these were cited by 8% or fewer 
respondents.

In contrast to what you often hear mentioned as a reason for choosing an 
open source database over a "big" enterprise database, 'Larger databases 
seemed like "overkill"' was only cited by 6.5% of respondents. Perhaps 
this is an indication of the types of users they surveyed.


The breakdown on usage for the various databases databases, according to 
the article (adds up to more than 100% as respondents apparently could 
choose multiple products):

74.7% SQL Server (down 1.7% from 76.4% a year ago)
54.5% Oracle (up 3.2%)
54.4% Microsoft Access (down 1.7%)
43.4% MySQL (up 4.9%)
23.5% IBM DB2 (up 3.1%)
11.2% PostgreSQL (down 0.4%)

To sift out the effect of legacy usage, the survey also asked "which 
databases were used for the most recently completed project." The 
results for that were:

51.0% SQL Server
37.1% Oracle
20.7% MySQL
14.9% Access
12.5% DB2
  4.2% PostgreSQL


This BZ Research study seem to be in contradiction of a Gartner study 
the same magazine referenced in the previous issue (page 43):
http://www.sdtimes.com/download/images/sdtimes178.pdf

where a chart shows for 2006:

47.1% Oracle
21.1% IBM
17.4% Microsoft
  7.9% Others
  3.2% Sybase
  3.2% Teradata

Though this is measuring market share, which isn't the same thing as usage.

Apparently all the open source offerings are lumped into "Others," if 
they're counted at all. Sometimes market share is measured strictly by 
sales revenue, which will obviously vastly under report open source 
usage. (This was just a small side-bar, and it doesn't mention how 
Gartner arrived at these figures. This is one of those typical teasers 
used to sell market reports by market research companies.)


If all of this talk of closed source databases topping the popularity 
charts got you down, there's always this other survey to consider:

http://www.sdtimes.com/article/LatestNews-20070801-19.html

   Survey: Developers Targeting Windows Less Often

   Surveys published by Evans Data over the past two years have
   concluded that there is an accelerating trend of developers
   abandoning Windows clients as target platforms, and Evans forecasts
   that the trend will continue in favor of developing for embedded
   platforms and Linux.

   ...development specifically for the Windows operating system has
   declined by 12 percent from one year ago, continuing a two-year
   slide.

   The survey found that the number of developers writing specifically
   for Linux had increased 34 percent, from 8.8 percent of respondents
   a year ago to 11.8 percent today. The data also shows that there was
   corresponding growth in development for embedded operating systems.
   ...
   Evans found that Windows use by developers themselves has remained
   steady.
   ...
   Microsoft declined to comment on the Evans survey. It is a client of
   Evans? and a member of its advisory board.
   ...
   Other findings from the survey show that JavaScript is the most
   widely used scripting language, eclipsing PHP, Python and Ruby.
   However, Ruby usage is expected to increase by half over a year?s
   time.

   Additionally, a third of North American developers are working with
   virtualization, and its adoption is expected to increase by 42.5
   percent within a year.


  -Tom

-- 
Tom Metro
Venture Logic, Newton, MA, USA
"Enterprise solutions through open source."
Professional Profile: http://tmetro.venturelogic.com/

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.