Lambda the Ultimate

If you click through the graph, you can type your own queries. Of the PL-related terms I tried, C and SQL had the biggest numbers, around 4% compared to ~.015% for the largest numbers on the original chart. Even RPG and Fortran are 3-6 times as big, in the .05% range.

Of course the orig ref is interested in the huge slope of the Groovy numbers, up from 0 to microscopic in a couple of years.

Actually, the graphic shows Groovy WITH Java combined...

A very important subtlety.

It is also funny it is Groovy and not JavaFX. Not like Sun intended! Oh well, Oracle's problem now.

Now if we could just juxtaposition that slope with "number of bugs caused due to Java Generics".

If this is obviously related to programming language research and theory

It's starting to look like we need a FAQ which reads "Is LtU devoted exclusively to programming language research and theory?" The answer would be "no".

The first paragraph of the FAQ should make this reasonably clear. If it really isn't, perhaps it needs to be updated. The first paragraph of the policies mentions "an emphasis" on research and theory - which we've actually considered watering down in response to the persistent misunderstanding about this.

The policies go on to point out that the discussion forum, where the current item was posted, is for "regular members to raise and discuss relevant topics that they think may be of interest to the LtU community."

So while I can think of any number of reasons to criticize this graph, not being on topic isn't one of them.

FAQ or not, it simply seemed out of character for the site. I've been here three years already; while not a long time relatively speaking, I'd hope it would be enough that I'd have a good feel for this sort of thing by now.

What I don't understand is why this is on topic, but something like this isn't. Despite the poster being whacked twice by the regulars here, some interesting discussion still started to form. (Perhaps there may have been more of it; with Ehud crashing in at the start, it's hard to say.) If anyone did actually have some insight into the practicality of the proposals given, I would've liked to have heard it. It's not even all that clear to me that the post violated policy (4a). If there are some brilliant "ideas that have been published and argued elsewhere" regarding how those three record systems compare, especially in terms of practicality of implementation, I'd appreciate a link.

Perhaps what's on and off topic for the site is obvious to many here, but it seems rather arbitrary to me. In particular, it seems that "the regulars" have a much greater tolerance shown to them when they post something that is of no clear interest to anyone whatsoever, much less LtU in particular. From my perspective, it would be more useful to have more posts like the one I linked to and less posts of useless graphs, Google Wave, and alternatives to Java for J2ME development. If I wanted that, I'd go to Reddit or Slashdot.

In short, I don't think we need yet another FAQ. I think we need a clear, concise, and objective policy first, followed by an enforcement of that policy with an eye towards not driving away new members and not stifling potentially interesting discussion. Most importantly, "the regulars" should be held to policy more strictly than new members, not the other way around.

FAQ or not, it simply seemed out of character for the site.

I'm not arguing that. As I said, I was responding on a specific point, that being related to programming language research and theory is not a requirement for a forum post or even a front page story. The reason I responded to that is because this seems to be a persistent misunderstanding, and the policies and appropriateness of items certainly aren't going to make sense if that's not understood.

What I don't understand is why this is on topic, but something like this isn't.

In both cases, "on topic" is not the issue. Posts like the one you linked to are discouraged as a matter of policy, as you alluded to. However, I think you may have misunderstood that policy:

It's not even all that clear to me that the post violated policy (4a). If there are some brilliant "ideas that have been published and argued elsewhere" regarding how those three record systems compare, especially in terms of practicality of implementation, I'd appreciate a link.

So would we, since that would follow the policy. The lack of such work doesn't imply that LtU is the place to hash out those ideas, however.

You apparently don't like this policy, but it was instituted for a reason. You can get some idea of the history of this by reading Blog not forum! and its comments.

Perhaps what's on and off topic for the site is obvious to many here, but it seems rather arbitrary to me.

Anyone can post a forum topic, and as such many forum topics are posted that are marginally appropriate at best. I suspect this contributes to the perception you're describing.

In particular, it seems that "the regulars" have a much greater tolerance shown to them when they post something that is of no clear interest to anyone whatsoever, much less LtU in particular.

In my opinion, that's a somewhat exaggerated perspective, perhaps fed by a few recent instances. However, the site intentionally caters to the more regular members of the community, i.e. people who post and comment over a period of time. They help determine what is and isn't appropriate, both positively by posting items and comments, and negatively by objecting to questionable posts.

Your original comment in this thread was an example of the latter, and that's as it should be — once again, the only reason I responded to it was to address the point about needing to be "related to programming language research and theory" to be on topic.

I'd say both this post and the How best to add a record type to my typed Scheme variant? post should be counted as clearly on topic: this post because it is talking about a PL usage trend for a class of PLs, and the record type post because it concerned the kind of language design issue that says something about the general phenomenon. I can't say about the Java and Google Wave stories, because the titles interested me too little for me to read them.

There's no doubt that regulars enjoy much greater tolerance than new members, but there's only one way to fix that: have other regulars, such as yourself, call them to account. So here I go — Ehud, David: I second John's complaint that Sam Rushing, the poster of the record type story, was unjustly criticised by you two. Policy 4(a) was applied too narrowly, without attention to what kind of discussion LtU should be encouraging.

And a point in defence of regulars: they are the ones who feel responsible for finding content to put up when the site is slow. Sometimes a little weak content is better than nothing at all, policy be damned. Furthermore, objectivity of policy is not a good in and of itself: what matters is reaching agreement on what is within bounds.

What I don't understand is why this is on topic, but something like this isn't.

With the particular post you mention, I think the problem is a historical one. The experience over the years has been that "help me with my language" postings tend to go off the rails.

That particular one turned out okay, but long-term members are "once bitten, twice shy", so I can understand the feeling.

One of LtU's challenges as a virtual community with self-appointed membership is that it is hard to get a feel for what the community is or should be. The best we can do is try to voice a consensus, and when that fails, the judgment of those few, such as Ehud and Anton, who have taken a greater responsibility for the care and feeding of LtU will have to be invoked.

Thanks Marc, I agree with all your points. I want to expand on the issue of avoiding "off the rails" posts:

That particular one turned out okay

It did, but it also contained a hint as to why such posts are discouraged. dmbarbour wrote in that thread, "There are plenty of times I'd like to ask similar questions..." This hints at a demand that is not easy to satisfy without compromising site quality.

From experience, we know that allowing such posts would result in a spectrum of quality, all the way down to beginner-level questions close to the "do my homework for me" level, which in LtU's case would mean that they'd be posted by people who've hardly made a dent in any of the standard intro-level material mentioned in the Getting Started thread.

Controlling that would require constant policing based on quality as opposed to type of content. That's the most difficult kind of policing to do, since (a) nobody likes to be told that what they posted was low quality, and (b) most members are understandably reluctant to do such policing themselves, and risk the resulting arguments. Ehud and I regularly receive private emails about such posts as a result. This either places a greater policing burden on the few members who have shown willingness to do it, or else means that more such posts go unremarked, making them appear acceptable and leading to further confusion and dilution about what is and isn't acceptable.

To relax the policies on such posts, it would help to have some sort of reasonably objective criteria by which to determine appropriateness. Those criteria would ideally need to be understandable by those tempted to post questionable questions.

One possibility would be to place a strong Policy #4 requirement on such posts, that any such question must include reference to a prior art search. For example, the post about typed records included references to work by Wand, Remy, and Ohori, demonstrating that the poster had done some research and was not a complete beginner trying to get others to do his work for him.

Pretty soon the forums will be flooded with PLT... people will be discussing stuff they're digesting at conferences they attend over the summer while they are away from their classrooms. Then August will be slow again like it is every year.

Regarding utterly useless graphs... what would you prefer to see to show programming language trends?

Suggestion for Ehud: throw some LtU data in a cube or data mining text analysis tool and see what people like to talk about most. It'd be cool if LtU's discussion served as data points to see what is hot in PLT research. Are you familiar with cubes and such, or does that sound intimidating? Also: Maybe an engineer at Google could assist LtU with such a project.

It was meant to be a community of like minded individuals. Like minded, that is, in the sense of being deeply interested in programming languages in the general sense of the term (including theory, but not limited to it). Part of being a community is getting to know other members.

I agree with this. But while healthy communities are more about their regulars than they are about short-stay visitors, equally they do need to be open to new members (cf. the pub scene at the beginning of American Werewolf in London), not least because they are where the regulars of the future come from. I'm bothered that "another rule for the regulars" is discouraging to just the kind of new members we should be encouraging. We should try to make the rule work for both regulars and visitors.

1. Sure.
2. There are no "rules". That was an intentional decision by the community. The policies document tries hard to make this clear, and it seems to me it does this reasonably well.