On Sun, Jul 01, 2001 at 11:33:44AM +0800, Selena Sol wrote:
> In other words, within the context of reading and thinking about 'this'
> list, I don't particularly care about "overly technical issues" such as why
> it is bad that Java does not work on a PowerPC with LINUX installed and how
> hard it is to install LWP.
>
> I don't care, because my clients (or the people who make technology
> purchasing decisions) don't care.
This is a valid POV. Just because someone has an AIX box (or any
other "boutique platform") doesn't invalidate their need to solve
problems though.
> Aren't we supposed to be working together to think of more efficient AND
> MORE REVENUE-GENERATING ways in which to advocate Perl here? (Remember, I
> am a newbie so I may be totally wrong).
There are projects out there where revenue generation isn't a primary
concern; cost to develop and time to develop predominate the decision
making, which may lead into revenue generation as a second-order effect.
I used to work for a company making enterprise software that had to be
truly cross platform. We used Perl; we deployed on multiple different
versions of AIX, Solaris, DEC Unix and Tru64. We would have left a few
million dollars on the table if we didn't have that flexibility.
> I believe that the the initial point made by Adam could be summarized as...
>
> "The Java development/application environment can be very difficult and time
> consuming to install."
>
> I think this is a valid point.
That's only half of it though: Perl isn't anywhere near as difficult
for your typical installation. Perl also is happily devoid of the
licensing issues, multiple implementation concerns, etc.
In all of the holy wars about Java vs. Perl, this is one point I've
never heard raised. Lots of times it comes down to disliking Sun or
lack of dynamic binding or deep object support -- issues that don't
have anything to do with the price of Tea in China.
> That said, a Perl install "of equivalent level complexity" is
> also pretty darn hard. Note that when I say equivalent
> complexity, I mean that a perl accelerator like mod_perl
> needs to be installed and a full-featured toolkit along
> with it. Only then does a perl solution compare with a
> servlet one.
OK. This is an important point. Probably something that can be
improved upon...
> Is our best advocacy strategy to say that, "we don't suck"? Or, more
> correctly, is our best advocacy strategy to say that our competition does
> suck? Isn't that really what the meaning behind the first email from Adam
> was?
It occured to me that with all of the bitching and moaning about Perl
these days, I stumbled across an aspect of Perl that's remarkably well
engineered and in fact doesn't suck. Many years of effort have gone
into the process, and today anything that can be made to look like a
UNIXish environment can install Perl pretty darn easily.
This is certainly not the case for Java2, outside the supported
platforms where point-and-click install binaries aren't available,
and source code isn't readily available since it's not open.
> Now, don't get me wrong, I do believe that slandering your competition is
> one of the tools in your marketing toolbox.
Slandering Java isn't the point. Highlighting Perl's strengths wrt Java
is the point. If Sun were to reform their support/licensing/distribution
policies, I'd be a much happier programmer.
> I think that Elaine is 100% correct in saying that the pointy-hairs will
> simply say that they can buy a support contract from SUN and the issue will
> disappear from their world.
I agree with Schwern in saying that's a cop-out. I don't need a support
contract to program in C, C++ or Perl, why *should* I need one for Java?
If I need to spend my ever-shrinking IT budget more wisely, is that
something I need more than an upgraded server or database engine?
> Furhter, any enterprise client will say that it
> does not make one bit of difference how hard it is to install Java on a
> LINUX PowerPC. That is not the infrastrucute they would be using anyway.
>
> They probably already use SUN/IBM/Microsoft or the like.
Is it the case that there is no problem, they don't see a problem (yet),
or that they don't care if a problem exists? Whichever way the enterprise
clients decide, their values and decisions are not universal.
Z.
