Comments on: A Modest Proposal

By: Stephen

Stephen — Fri, 13 May 2011 11:18:26 +0000

Erg.. major type-o — I meant to say, “I can’t believe there isn’t more of a movement”

By: Stephen

Stephen — Fri, 13 May 2011 11:17:57 +0000

I can’t believe there is more of a movement out there to make Clojure more designer-friendly. Sad to see the project didn’t take off in the past couple years, but I’m hoping folks will start to realize that this kind of thing is needed.

The project was a great start, though! If more effort gets put into it, I will definitely try it on some big projects.

By: Shantanu Kumar

Shantanu Kumar — Wed, 01 Sep 2010 18:04:13 +0000

@tim This is great. I would suggest that you separate this out from compojure and servlet stuff. The reasons are:
1. People can use it for template processing and offline doc generation.
2. Web frameworks other than Compojure can use it, thus fueling innovation.

I would be eager to know when/if you factor this out into a compojure/servlet independent template library.

By: Corwin Jensen

Corwin Jensen — Tue, 15 Jun 2010 17:06:25 +0000

I think that while a good segment of the market still embraces the PHP way, we’re seeing a migration away from this type of development (even in the PHP world). I don’t think that reinventing it in any language is a good way to promote that language, especially when the language itself tend to attract the more esoteric crowd. It’s a bit like putting sweats on Michelangelo’s David. You might attract a certain type, but you’ll probably alienate your real target audience. I was personally attracted to Clojure for the following reasons:
+1

By: John Hunter

John Hunter — Sat, 06 Feb 2010 02:53:15 +0000

I think this is a great idea !
There is something satisfying about getting instant feedback as in php.
There’s no reason something like this “Clojure Server Pages” and compojure couldn’t both be useful.
Maybe you begin with this and later head into compojure.
Did anything ever happen with this, Tim?

Cheers,
John

By: Dan

Dan — Tue, 01 Dec 2009 19:38:46 +0000

I think the most fundamental place where Clojure can make a serious contribution is in the sciences. Currently, code is not treated as data in the sciences, and this is a problem. Much of the new code in my area (machine learning / neuroscience) is written in Python, where code is most definitely not data.

This makes it incredibly difficult to write programs that can keep track of how different data sets were arrived at. When I transform a data set in some way (usually a whole series of them), I want my operations to be published with the new data set. I want them to be published in a format that is executable, easily legible, and most importantly easily parsable – I want to be able to use other code to reason about these transfomations, try new ones, etc. etc.

In other words, I want to use the code as data. Clojure (especially with Incanter), offers the possibility of doing this. I can write programs that compose programs, reason about them, search through them. I can use the same language to write and read each layer of program. I can fairly easily distribute my programs over the cluster thanks to integration with Cascading and Hadoop.

By: Greg

Greg — Fri, 09 Oct 2009 20:08:50 +0000

BTW, I can’t believe you’re getting naysayers, this sort of stuff is essential for making great dynamic website (IMO), especially when *designers* are making website — they’re used to the PHP-way of doing things, and personally I don’t see any problem mixing HTML with another language, I don’t think it’s ugly at all, but as Cliff pointed out, there are plenty of people who do.

To each his own, thanks again!

By: Greg

Greg — Fri, 09 Oct 2009 20:03:56 +0000

Tim, thank you for doing this, I was afraid I’d have to do this from scratch myself, but now that you’ve done this it makes things much easier, *and* I’d be happy to contribute to the project as well. I completely agree with your conclusion, we need a great PHP-style system to compliment Compojure, and this seems to be it (unless there have been other attempts that I’m not aware of).

I look forward to how this progresses, great job!

By: Cliff Wells

Cliff Wells — Sat, 12 Sep 2009 02:54:33 +0000

Your arguments echo many I’ve heard in the past with regard to Python. In fact, it was possible, using mod_python and some horrible PHP-like languages to get pretty close to the PHP-experience”. In the end, almost everyone rejected these approaches as they create horrible-looking mish-mashes of the two languages (yes, I consider HTML a language).

1) functional language features
2) comprehensive and modern libraries, thanks to Java integration (but without the pain of Java)
3) modern MVC web framework being developed using it (arguably a sub-category of (2) but important enough in its own right to qualify as a separate point)
4) S-expression based HTML generation (yes, I much prefer this approach, see http://breve.twisty-industries.com for my previous attempt to bring this type of functionality to Python).

We’ve seen widespread migration (largely thanks to Rails) away from the PHP way. I don’t think there’s a future in it and it appears the world at large is starting to agree.

By: tim

tim — Tue, 01 Sep 2009 21:29:37 +0000

@Antti: I hope you’re right — I really think highly of the language. It just seems like Scala has more momentum in the shops that are already using Java; it’s possible that Clojure could find success outside of established Java shops, but it’s a harder sell, maybe.

It’s possible this whole exercise might be a bad idea — it’s kind of a work in progress — but nonetheless I think that making Clojure web apps drop-dead simple is not inherently bad. Yes, there are 300 advantages to clojure.html, but it looks scary to developers: “What, I don’t even get to use HTML, I have to use another language?” So why not make it more approachable?

By: Antti Rasinen

Antti Rasinen — Tue, 01 Sep 2009 20:25:22 +0000

I think Clojure is already “made it”. It seems modern and fresh in the sense Arc failed to be. Clojure has the potential to become the biggest Lisp around. That is an achievement in itself.

To make it in the real world is of course a different matter. There may be truth in the statement that Clojure needs a niche. PHP is a good analogy. Other “Worse is better” examples apply too. But the rest of your article seems to describe PHP made out of Lisp? That is such a bad idea on at least three levels.

1. Is HTML templating truly an area where Clojure could shine?

2. Web frameworks and HTML templating are no niches anymore. These days you can’t swing a dead cat without hitting a web framework. Heck, you can’t even pick up the cat.

3. (Related to #1) Trying to ape another API or programming model in a whole different programming paradigm will lead to poor implementations and un-idiomic code.

It’d be great to see a popular Lisp and I hope Clojure makes it. But I do hope it makes it on its own merits.

I’ll make one prediction here. The niche, whatever it will be, will be a field where current programmatic solutions involve lot of repetition or drudge work. The Clojure/Lisp solution will remove that drudge. (Think Rails and CRUD, for example.)