[Haskell-cafe] Re: Libraries for Commercial Users

[Iain Barnett]

On 25 Oct 2009, at 15:51, Curt Sampson wrote:
>
> Funny, I do too. Still, when Luxuria opened for them in Vancouver  
> in the
> '90s, I started to think about what Howard Devoto was doing....

Anyone who recognises The Mark E. Smith is worthy of being master of  
all he surveys.

>
> But in others, such as the Viaweb one, you'll lose, as we've seen.

I don't really mean to say that it's a Ford production line, but you  
have to take into account risks, and staff leaving that you can't  
replace is a big risk.

The Viaweb thing is interesting, but I'm never sure how much it  
really means. Paul Graham is obviously a talented developer, but was  
it also a big slice of luck too? How many other talented Lisp/ 
whatever developers had their startups fail? It's not like people to  
write a series of articles about how great they were but they failed,  
there might be 1000 failed Lisp devs who never fessed up to their  
websites being crap? Likelihood is there was a mixture of luck and  
talent, and the all round implementation of his site was better.  
Facebook is PHP, Twitter is Ruby/Scala - it's not just the talent of  
the developers or the language that makes success (obviously).


On 25 Oct 2009, at 16:39, Magnus Therning wrote:
> Again based on personal experience it's not uncommon to
> decide to use a pre-existing library, and then half-way through  
> find that
> there are some serious limitations that only manifested after heavy  
> usage.
> Prompting a rewrite in the end.

I agree, that happens a lot. But these libraries do get you started  
very quickly, and that's extremely important. Better to be 90% there  
and struggle with the 10%, because that 90 buys you a lot of time.  
Having 10% that's perfect and the rest "is on the way and will be  
perfect but the site isn't up yet" won't inspire a customer.


What this discussion appears to come down to (in my opinion), is that  
there are 2 sets of developers in Haskell. You're part of the  
talented bunch that know it well enough to cook up your own web  
framework, or whatever, even when the libraries are missing. For the  
lesser lights of this world like myself, there needs to be some  
libraries to just get things going. I may never understand how that  
library is written - it doesn't matter (to me, and why should it? I  
thought that's what interfaces are for) That's what PHP has in  
abundance.

What it also has, and Ruby does very well too, and (in my view)  
Haskell doesn't, is lots of good documentation. Again, you might be  
one of the talented crew who see a new library come out with the  
barely obligatory one line description on Hackage, and all you need  
is the API docs and you're away. That won't cut it for the majority.  
There need to be tutorials, examples, blogs for the masses -  
everything a "big" language has.

It looks like a bit of a decision for those who are already happy  
with Haskell, are they willing to make it more accessible to others?  
(which is really a euphemism for "make it more commercial") They  
don't have to. As you say, you can write anything you need bespoke.  
It's the way the Perl mongers like it, and it's not gaining any  
ground, but do they care?

But Haskell is far more beautiful, so I'd say give the people (me)  
what they want.

Iain


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.haskell.org/pipermail/haskell-cafe/attachments/20091027/61b5b4d0/attachment.html