Hacker News new | past | comments | ask | show | jobs | submit login
The Haskell Platform: a single, standard Haskell distribution for every system (haskell.org)
34 points by jlhamilton on May 6, 2009 | hide | past | favorite | 18 comments



No DB interface included, not even ODBC. Team Haskell doesn't seem to understand (or care) what a barrier this is to mainstream adoption.

(Yes I know about HSQL but it's not exactly "batteries included" is it - compare Haskell to Python in this case)


We didn't want to make any controversial decisions on behalf of the community in the first release. For subsequent major releases will will have a community-based decision procedure for what new stuff to add. See http://blog.well-typed.com/2009/05/next-steps-for-the-haskel...

You'd have to lobby for say, sqlite or hdbc or gtk2hs etc. Snoop around on hackage.


lol pwnd


"Future releases of the platform will include more tools and libraries to meet developer need."

In other words, tell the Platform team it's needed, and get other people to do the same. Even better, write the database interface you'd like to use and push them to adopt it.


The single most important element of this package is the cabal-install tool. "cabal install HDBC" or "cabal install hsql" etc. will download, compile and install any library on hackage. Previously you had to 'bootstrap' cabal-install yourself, which wasn't a big deal for experienced users but was an annoyingly significant barrier for newcomers.


While I agree unreservedly with this, it's probably worth pointing out for the benefit of those who haven't gone through the process themselves that bootstrapping cabal-install is a five-minute job (or less, depending on how many dependencies you need to compile). The barrier is that one needs to look up the instructions, and know that one should be looking looking something up, rather than it being in any way difficult or complex.

http://haskell.org/haskellwiki/Cabal-Install


I appreciate that; I am something of a "languages pioneer" at work, and I have gotten good traction with Python because the learning curve is very shallow. You can use it a a glorified shell script and get something useful done very quickly and as you learn more you can add in the FP goodness, OO, whatever you want, it grows with you. Now looking to take it a bit further, I am shopping around, Haskell and OCaml are candidates, but to sell this on the grounds of, learning this and using it will get your project done quicker than continuing in what you're using now. So I have shot myself in the foot a bit; Python is really productive, and other than the language itself, the biggest factor is the ActiveState "batteries included" distribution.


This install 6.10.2, right?

I grabbed the generic source installer bundle.

When I run ./configure, it tells me

configure: error: The Haskell Platform version 2009.2.0 requires ghc-6.10.2

The requirement to have a working GHC in order to get a working GHC, while interesting, seems a perplexing roadblock.

But extra puzzling is the requirement to have already the same version of GHC that this is supposed to be installing.

What sense does this make?

I really like Haskell from what I've seen so far, but every time I try to explore something that requires more than one or two packages it feels like a test of cunning, will power, and patience to get all the right pieces and version in place.

(BTW, I have GHC 6.11.20090501 installed via darcs source. Apparently this is not enough:

./configure --enable-unsupported-ghc-version ... [much stuff] ...

checking for GL/glut.h... no configure: error: The GLUT C library is required

:(

I appreciate the effort and the ambition, but it's still too tricky.


Clever, clever logo :)


> The Haskell Platform is a blessed library and tool suite for Haskell, ....

Can someone explain what "blessed" means, in this context?


now... lisp.. if only


Off topic question:

Can somebody explain to me whether there is a semantic difference between "..." and ".." above? I find this a baffling construction, moreso because both happen within the span of two words, but at the same time it seems unimaginable that it is an unintentional omission, because the cognitive load of typing out "..." is a one-time retrieval + motor command and should be automatized.

On the other hand, if the omission is a mistake then it means that it isn't automatized and there should be a reason for that as well. It is also possible that it is actually a formal punctuation mark in another language, which leads to comparable and competing access times with "...", in which case I have no knowledge.

If the ".." has indeed acquired actual, independent semantic significance it must be recent and I cannot find it; what would it be?


I read it as a trailing thought. Keyboards don't give us much expressiveness, but we make do. :P


I see. Is this the commonly-accepted interpretation? What does the normal ellipsis mean then? I use it as a trailing thought as well -- or, is it trailing speech?

    spoken: now... [pause]
    spoken: lisp .. [thought...?]
    thought: if only

?


wow. I never thought a comment of mine would be tokenized.

But, I guess your parsing was spot on

...


[Edit: I'm an idiot. This is a "batteries included" Haskell. If only Lisp, indeed]

CLISP + Emacs and slime is pretty much universal, almost to a fault.


Wait, who uses CLISP? Most Lisp applications included with my OS use GCL, and I personally use SBCL. It also seems like most libraries support SBCL, sometimes exclusively.


.. on win32?

I use sbcl on win32 and it does most things almost right. Just not all. At least CLISP doesn't half-run them.




Consider applying for YC's Summer 2025 batch! Applications are open till May 13

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: