programming contests as a way of finding/evaluating offshore talent...

[Patrick J. Kobly]

Quoting PFJ <paul at all-the-johnsons.co.uk>:

> Hi,
>
>> i'm looking for opinions on the "worth" of programming contests as a way of
>> judging the talent of potential software developers...
> Cons

To the list of cons, I'd add:

- Most programming contests are not representative of the work developers do
day-to-day.  In particular, algorithm competitions are not at all
representative.  Topcoder's component design/dev/integration competitions may
be more applicable but YMMV.  Just make sure the process implied by the
competition is comparable to the process you use for work.  Further,
competitions do not evaluate soft skills like teamwork, requirements
elicitation and negotiation that professional developers use on a regular
basis.
- Documentation (user and developer doco) is almost never evaluated.
- Testing skills are rarely evaluated.  This is obviously relevant even 
if your
shop has a strong, separate QA department.
- Contests often limit tool-support to a specific set of compilers, languages
and IDE's.  Often, tool selection (and adaptability) is a critical part of a
professional developer's expertise that is not tested by these contests.
- Time constraints measured in hours (as with many competitions, exceptions
being TCO's component and marathon matches) do not adequately test time
management skills.  When a developer is applied toward a more complex total
problem, with a longer time-frame, you don't know how they will deal with
deadlines.
- Contestants are not evaluated in a similar manner to how they are 
evaluated in
a professional development position.  As a result, they code to 
evaluation - if
they are evaluated only on algorithm correctness and speed, they code for
algorithm correctness and speed, ignoring readability, robustness, portability
and maintainability.  This is not necessarily indicative of how they would
perform professionally.

Many of these concerns can be addressed if you conduct the competition 
(or have
influence into how it will be conducted), but you have to be aware of what is
being tested and what aspects of a development position are being ignored.
Also, if you are conducting the contests (instead of using a third party with
an existing ccompetitor base), remember that the "find" part of your goal will
be negatively impacted.

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: PGP Digital Signature
URL: <http://listman.redhat.com/archives/fedora-list/attachments/20060419/71298900/attachment-0001.sig>