Front page | perl.perl5.porters |
Postings from August 2001
Re: The case for SDKs
From: Peter Scott
August 3, 2001 13:16
Re: The case for SDKs
Message ID: email@example.com
At 02:51 PM 8/3/01 -0500, Elaine -HFB- Ashton wrote:
>Peter Scott [Peter@PSDT.com] quoth:
>*>As I recall the meeting, the primary rationale for SDKs is to reduce the
>*>number of things that Perl SAs - the example was an ISP/ASP - have to think
>*>about when installing software for their customers. Which to me is a
>*>compelling argument. So they only have to think, "We've got Perl 5.6.1,
>*>the Perl Web Services SDK 1.17, the Perl Database Services SDK 2.3, and the
>*>Perl XML SDK 3.4, we're wearing sunglasses and it's 106 miles to Chicago" -
>*>oops :-) Anyway, the point is that they don't have to think "Perl 5.6.1,
>*>plus about 45 other modules whose volume numbers have to be obtained from
>*>the source or the tar file name".
>See, I get grumpy when people put 'SA doesn't have to think' into a
>sentence. Not that I don't appreciate the sentiment of making thing easy,
>but having the SA conscious during some point of their career is a bonus.
Eh, you know what I mean. They have enough things to think about
already. I do enough of that kind of work to feel their pain :-)
>If re-doing the CPAN FAQ has taught me anything is that you can answer a
>question, you can answer it several ways in complete detail, and you can
>add the <blink> tag and you know what, the idiopts your are trying to help
>are *still* going to send you email saying they have seached high and low
>looking for the answer to this horribly difficult question. :)
Oh, I don't give a fig about the idiots. I'm on the "make easy things
>The sentiment and the goal you seek is noble and good but it isn't going
>to help the people you really want to help.
>*>To satisfy this rationale we merely need to build SDKs that contain the
>*>'popular' modules that these providers' customers want. We don't have to
>*>make all the modules sing in tune.
>Everyone has their own idea of what an SDK might hold or be or do and
>after two years watching tennis on the SDK list, I'm really high on the
>'try the bundles' idea since it's pretty clear that a) getting agreement
>on the contents b) getting it built on 6 or 8 major platforms and c)
>keeping it updated is going to be difficult if not nearly impossible and
>even then, it's not a near future thing.
>Try the bundles, love the bundles :)
I think the SDKs *are* bundles, just with explicit versions. But that they
should be blessed by a pumpking, otherwise we're making work for people who
have to ask, "Which is best, Bundle::Web::HFB, Bundle::Web::MSERGEANT,
Bundle::Web::MERLYN, or Bundle::Web::MWRIGHT???"
Pacific Systems Design Technologies