Home » forums » Site Discussion

How is LtU doing?

I thought it was just summer, but now that fall is here, the site still feels kind of quiet.

By Sean McDirmid at 2013-09-12 06:54 | Site Discussion | previous forum topic | next forum topic | other blogs | 10000 reads

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Work work work

Working a day job taps my motivation. I try to spend what little remains on Kitten, preparing it for release into the wild. Not much time to make interesting posts.

By Jon Purdy at Fri, 2013-09-13 17:48 | login or register to post comments

Kitten sounds like a great

Kitten sounds like a great PL name.

I'm just afraid LtU is losing out to the new stuff, but it would be a shame because google plus and hackernews are so noisy with other stuff.

By Sean McDirmid at Sat, 2013-09-21 12:00 | login or register to post comments

I still check here regularly

It’s just a slower-paced community. There’s quite a bit of value in the archives, and it takes time to consume articles and papers. For example, I’m still going through “Understanding and Expressing Scalable Concurrency” and mean to post my review once finished.

By Jon Purdy at Sun, 2013-09-22 12:44 | login or register to post comments

Want me to start a new discussion?

I only check the site every two or three days now, instead of hours. (I use googleplus not at all, and hackernews has—in the last year—had a dramatic increase of inane yammering so I read it less.)

I'm designing something, if you want to talk about it in a new discussion topic. The subject might be "lightweight process models", and goes so far as to define basics in order to reground criteria of analysis when it comes to stuff we often take for granted. For example, it becomes necessary to define what verb "to happen" means to clarify what sort of evidence can be generated about it.

I haven't written about it elsewhere, so that breaks rule number one. Also, it has a pretty strong focus on practice as opposed to theory. But I think parts will strike folks as theory: the semantic model parts, which I find interesting.

Lately I've been thinking about defining parts of this computation model all the time when I'm not at work. Instead of playing video games, I watch my son play on the Xbox and think about this computation model instead of paying attention to his game. Part of the process involves trying out language describing the model, to see where it goes wrong by implying lines of reasoning that aren't fitting.

That's why I want to discuss it, to try out phrasing on a real audience. But I'm more inclined to write code than write about it, these days anyway. It's tempting to quit my current job and work on it privately for a year, but job hunting again when you're not already working is quite unpleasant. So I'd rather avoid that.

By Rys McCusker at Tue, 2013-09-24 21:10 | login or register to post comments

Frontpage editor?

I think I would possibly propose more stuff for discussion if I could submit things on the frontpage -- this has been raised recently. I'm not exactly sure why there is this distinction on LtU anyway.

By gasche at Thu, 2013-09-26 02:26 | login or register to post comments

Is the LtU front page

Is the LtU front page related to upvoting content on hackernews/reddit? Does it make sense to rely on editors in LtU, or to just add an upvote mechanism somehow so that popular content floats to the top?

By Sean McDirmid at Thu, 2013-09-26 03:41 | login or register to post comments

Drop Ehud an email

and you're a frontpage editor in a minute: http://ehudlamm.com/

By Manuel J. Simoni at Thu, 2013-09-26 12:45 | login or register to post comments

Done!

Done!

By Ehud Lamm at Sat, 2013-09-28 08:15 | login or register to post comments

Good idea

Thanks to your suggestion and Ehud prompt answer, I just posted my first "front page story". I went from RobJellinghaus's implicit suggestion that we could discuss this year's ICFP papers and posted the previous work of one of them.

I'm considering posting a handful more stories about stuff I appreciated at this year's ICFP (and workshops), but I'm not sure what a good rythm is and how much is too many -- the point is not to grab the mike and flood everyone with my own pet peeve, type systems for programming language design. I've been thinking once a week or something, feedback appreciated.

By gasche at Sun, 2013-09-29 23:55 | login or register to post comments

The main reason is

The main reason is historical. This is how the Manila software we originally used worked. But it served us well, so I never felt it should be changed. Anyone interested in PL and who follows the site can easily become a contributing editor.

By Ehud Lamm at Sat, 2013-09-28 08:20 | login or register to post comments

Seems to be a lot going on out there

I'm also a bit bemused by the relative quiet here, given that from my perspective there has seldom been so much language innovation. Chisel and generally the entire emerging Scala DSL ecology; Ceylon created by Gavin King, formerly of Hibernate and Seam; dependent type systems busting out all over, including one that may wind up in the Unreal game engine, and another aimed at typing jQuery....

I am not sure whether the silence here is because the language discussion is happening elsewhere, or because a lot of this work is taking research that LtU has known about for a long time and trying to bring it into practice, or what. I feel like I could singlehandedly populate the front page.

But really I satiate most of my language jones just through the ACM digital library. In fact, this weekend is one of my favorite times of the year, with ICFP and the epic pile of associated workshops (Haskell, Erlang, dependent typing, etc., etc.) all putting up their actual papers at once. Bonanza!!! It would be interesting to discuss just about every paper published in that whole conference-plex....

By RobJellinghaus at Wed, 2013-09-25 16:35 | login or register to post comments

dependent type systems

dependent type systems busting out all over, including one that may wind up in the Unreal game engine

Interesting that they don't cite this prior work, which seem to implement what they're after, ie. partial type checking for an undecidable pure type system.

By naasking at Wed, 2013-09-25 19:13 | login or register to post comments

What works?

A recent thread The Three Laws of Programming Language Design is up to 80 thousand, plus hits. I am amazed by this. The average is only 10 to 20 thousand. What is it saying that is so important?

By Hank Thediek at Thu, 2013-09-26 10:40 | login or register to post comments

Maybe entertaining rather than important.

Some traffic was due to a HackerNews item hitting the front page with a link to LtU. Also, Erlang is a popular language and Joe Armstrong is well-liked, and says intelligent things. This time Joe complained about behavior of other folks in their reception of Erlang detail. (I know, complaining about complaining is very meta. But many online conversations are basically circle put-downs, which makes them boring.)

A sub-thread on monads might have given the item longer legs. Rather than importance, I think the draw was whatever folks like reading (Armstrong, dissing, and monads?) plus drive-by gawking.

Though a possibly dubious goal, If you want traffic, I think what works is material folks think might influence what they're doing now, or expect to do later. Tech folks have a taste for insight that can inform future planning. A post on Golang will have more hits than one about types, if folks wonder about Go and try to avoid dry type theory stuff.

By Rys McCusker at Thu, 2013-09-26 23:06 | login or register to post comments

On the evening news, you see

On the evening news, you see some news that is useful, and some news that is fluff. The fluff is necessary to keep the viewers coming back, hopefully to see news that is useful. LtU could use some more fluff to grow its user base and become a nice sounding board for PL ideas and news. (the same is true in conferences and so on)

By Sean McDirmid at Fri, 2013-09-27 23:48 | login or register to post comments

The problem of LtU is not

The problem of LtU today is not that it has too little fluff, but that it has too much fluff.

By Andreas Rossberg at Sat, 2013-09-28 14:40 | login or register to post comments

Fluff can go elsewhere

I think content that we really think interesting is enough to attract readers if we take the habit of linking good discussions on sites that already have fluff -- eg. reddit. Ensuring that people see value in the discussions happening here is, I think, the best way to keep them happening (and heard).

By gasche at Sat, 2013-09-28 15:49 | login or register to post comments

Fluff ratio? And recycliing

I agree with Andreas and gasche here that enough fluff can be found elsewhere. I think a little bit is useful, though. Keeping a ratio in policy, like 1:4 fluff:crunch, might enable a good balance.

I'd like to see old-but-great papers brought to the front page more often, i.e. sort of 'blast from the past', or 'then and now' styles. My intuition is that these should be about half of LtU's content; there are so many old, good, mostly forgotten papers, and fresh discussion (or at least reminders) of them could be interesting.

By dmbarbour at Sat, 2013-09-28 16:45 | login or register to post comments

gateway drugs

We will kill our community if we completely avoid marketing and populism. I'm finding it hard to connect with the reddit and HN communities that include a lot of PL enthusiasts, but who don't bother with our high-brow PDFs and are consuming their content online in much softer forms (e.g. shorter web essays). We become boring and disconnected, because that is the only face we show them, where a bit of fun and excitement could go a long way in making sure more serious formal material at least has the chance of being seen.

Actually, modern relevance is a huge problem with the PL community at large; our papers are increasingly abstract and their motivation is buried...either it does not really exist, isn't seen as very important, or is part of a peer review echo chamber. Some amount of connection with a broader community would better help us stay relevant.

By Sean McDirmid at Sun, 2013-09-29 00:16 | login or register to post comments