[Catalyst] Re: Catalyst Digest, Vol 10, Issue 91

[Charlton Wilbur]

On Dec 24, 2005, at 3:04 PM, mbailey at vortexit.net wrote:

> This gripe apparently seems born out of frustrations with your  
> client, but
> you're not alone.  Clients wanting the world, wanting it built  
> yesterday,
> and for no cost, is a business mentality; it will still be around long
> after both you and I are dead.  I believe saying you can have
> authorization and the multi-tude of other "bitch-work" things done for
> your app in "short time" IS A GOOD sell for Catalyst or any framework.
> CPAN, Not being tied to the web, and object-oriented development 
> (and it's
> associated unit testing) are all things that can be accomplished  
> WITHOUT
> Catalyst, so I don't feel those are strong selling points for it.  
> It is
> the speed at which these things can be accomplished, that's the  
> selling
> point.  So why not emphasize it?

Oh, emphasize it, by all means, just don't do it hyperbolically.

"You'll get 80% of your application done in an hour!" is an obvious  
lie.  It really only applies if the application is a close match to  
something that's provided as an example, or if the application lines  
up neatly with the scaffolding provided, and even then "an hour" is  
an optimistic estimate for someone learning a new framework.  The  
first gotcha with Apache or with Perl or with the operating system  
will also pretty much demolish the hour timeframe, and there are a  
*lot* of gotchas there; those of us who are used to dealing with the  
web already know where they are and don't get caught up on them (or  
at least know to check them first), but the new programmer that this  
marketing hype is aimed at won't.

More to the point, CIOs are used to this sort of lie, and see right  
through it.  They *know* that 80% of the application won't be done in  
an hour, and without any other emphasized selling points, why use  
Catalyst over J2EE or even PHP?  I mean, PHP lets you get something  
up on the web in a matter of minutes, too, and if there's nothing to  
allow a CIO who last touched code before the web to evaluate the  
merits of Catalyst beyond a hyperbolic statement about development  
time, then how on earth do we expect anyone to adopt Catalyst on a  
broad scale?

The other three points are not Catalyst-specific, but they offer much  
more compelling reasons to choose Catalyst than development time:   
flexibility and maintenance time.  It really doesn't matter if the  
program takes 20% of the time to write if it takes 150% as long to  
debug and 300% as long to modify later, because that's where you  
spend most of your time.  And that's where CPAN and OO unit tests  
become relevant, and that's where smart CIOs know they can save money  
and time.

> btw.  To catch the attention of people who aren't working on small  
> apps,
> you need to catch their CIO's and CIO's are interested in lowest  
> common
> denominators, how cheap can I get someone in here to work on this 
> (ie. Good
> Documentation and not complex Design), and how much time(ie. MONEY)  
> will
> this save me on X project.  Oh..and once this is written how  
> difficult(ie.
> Expensive) will it be to maintain this over X technology(ie. how  
> stable is
> it? how long has it been around? can I pay someone to fix it, if I  
> can't
> fix it myself? how much will that cost?).

Exactly.  Right now Catalyst, as useful as it is, is being *marketed*  
to hobbyist programmers and programmers with a bit of spare time who  
have a little project they want done.  There's nothing wrong with  
that, but if you want things like O'Reilly books there needs to be a  
market for them, and that market is largely companies with book- 
buying budgets.

And *maintenance cost* -- which means automated testing and code  
reuse -- is critical to this market.  Sure, CPAN and automated  
testing are not exclusive to Catalyst, but pushing them to this  
market will get Catalyst adopted a whole lot faster than pushing "80%  
of your application within an hour!"

Charlton




-- 
Charlton Wilbur
cwilbur at tortus.com