From: Joseph <jlhurtado@...>
Date: 2006-09-10T02:40:18+09:00
Subject: Re: Joel Spolsky on languages for web programming

Folks,

Well this is a LONG discussion!

Some points I wanted to address that have been raised by different
posters, I have included a short version of their own post below:

VIDAR HOKSTAD SAID
Sometimes the payoff in trying a technology that your team is
inexperienced with or that isn't widely deployed ...or the risks are
mitigated by your teams experience (Vidar has an application that is
already written in another language)

PETER BOOTH SAID
One point he doesn't make is the difference OPM and MHC (Other Peoples
Money versus My Hard-earned Cash.) When you are spending OPM costs can
become unreal. So as an employee a "buzz-word compliant/standard
approach" Java or NET solution might make more sense...


RESPONSE
Vidar and Peter I agree partially with your main points, however I
would argue your points apply to existing applications, software
already written in a given language where migration risks may truly
outweight embracing a new technology, and then perhaps staying where
you are is a safer, wiser choice.

I say perhaps because:
- Your code base would have to be significant, and your current
language at least good enough to achieve your business or startup
objectives in a reasonable amount of time.
- The productivity savings of the alternative (Ruby and RoR) would have
to be low or unimportant enough to ignore the other language.
- Other people's money truly dislike risk, any risk, and of course if
you have convinced them that
[insert-your-safe-language-framework-combination-here] is the best, and
the coolest, safest tools to use... how would you go back to them,
asking for money and telling them you were wrong... tough to do, yes,
but on ocassion I would argue WISE.  Because if this savings in
productivity were coming to benefit YOUR HARD EARNED CASH (MHC) then I
could bet Ruby would be the only obvious answer!

Also, a little off topic, you could reach a middle of the ground
compromise, extending your app in Ruby and leaving the "legacy" code
alone, a common practice in the mainframe world where COBOL apps have
been largely left in the cold or with minimal maintenace, while hooks
have been made into Java based applications that extend them!  Why not
do the same with Net/Java apss with links to Ruby on Rails?

Finally to end my reply I would say that if you are starting a NEW
project, mission critical or not, and there is no "showstopper" for
your application, by all means Ruby and Ruby on Rails are indeed the
best possible tools you could choose.  THAT IS PRECISELY my point.  The
productivity savings, the elegance and clarity of the framework is just
to good to ignore.

Best Regards,

Jose L. Hurtado
Web Developer
Toronto, Canada