Re: [boost] Review: lambda library

From: joel de guzman (djowel_at_[hidden])
Date: 2002-03-21 21:27:51

----- Original Message -----
From: "Jeremy Siek" :

>
> I'd encourage people to not vote against the lambda library simply because
> there are "promising alternatives". If these alternatives are for real,
> then the authors can submit those libraries to Boost when they are ready.
> If they end up being better, then LL can be deprecated, or they can
> coexist and users can make up their own mind.
>
> Further, many of the differences between lambda and Pheonix look like
> trivial syntactic differences to me, brackets instead of parenthesis, etc.
> With previous library reviews, the review manager has made suggestions as
> to which of such options were more preferable, and then the library author
> took this into account in the final polish-up before checkin. I don't see
> why that approach shouldn't be also taken in this situation.

Beyond the syntax, extensibility, IMO, is most important, especially for
someone who will want to incorporate such a library inside another
library. For instance, I was left scratching my head how to incorporate
local variables and closures to LL. With Phoenix, it took just a couple of
hours of coding to get this extensions. I think LL caters well to end users.
Users who will want to use it as an alternative to stl's binders, for example.
Beyond that, an FP library in C++ should also cater to other library writers
like me who would want to incorporate such a library inside another library.
For that matter, extensibility is a prime concern. There are lots of libraries
out there that will benefit from an FP library. Spirit's semantic actions is
an example. The signal-slots and other callback libraries also come to
mind. GUI library event handling code. Etc. While LL is very feature rich,
incorporating everything but the kitchen sink simply won't cut it.

That said, my opinion is that LL *should be accepted* to boost in its
current form here and now. LL fills a need. The FP paradigm is such
a powerful force that many could benefit from. It would do us all good
to have an FP lib in boost right now. Taking that further, I have such
high respects to these guys. The pioneering works of Jaakko Järvi and
Gary Powell (and Yannis Smaragdakis and Brian McNamara) were
the main influences of Phoenix. Quoting David Abrahams "The recent
developments with Spirit (Phoenix) and FACT! would probably never have
occurred without the groundwork laid by the LL". For Phoenix, I'll change
the word "probably" to "definitely".

LL is already a mature and stable library. I vote that LL be accepted to boost.

Regards,
--Joel