Earlier this week, I’ve been listening to Episode #87 of the (then) Stack Overflow Podcast, where Jeff Atwood and Joel Spolsky as well as the (then) new Community Coordinator Robert Cartaino discussed some of the fundamental features of and ideas behind the Stack Exchange platform, which sets it apart from other interactive community platforms such as phpBB or vBulletin.
Websites built on Stack Exchange (SE) are Q&A sites. And while some of the other community platforms (e.g., see above) are often used for this kind of interaction – people asking questions and receiving answers -, there are some features that do set SE apart from forum/message board (MB) type platforms. In the following, I will sum up some of those features. While the (incomprehensive) list is inspired by the podcast, it does reflect my own opinions only … one of the latter being that SE is great technology.
1. SE’s focus on answers is much stronger than MB’s.
MBs are general type discussion sites. They do allow for Q&A type content, but do not restrict content to be of a Q&A nature. They allow for discussion, opinion, off-topic content. SE restricts content to questions and answers. It does allow for comments, but answers are the strongest currency on SE-based sites, and comments exist on a much lower level only.
The structure of an SE page for a single answer allows for optimal visibility of the actual answer to the question! The answer that has been voted best (for a discussion of site moderation see section below) or accepted will appear just below the answer. On an MB site you will have to scroll down/click through to the last (or one to last) message in the thread to find the answer at best, or read through the whole thread in a worst case scenario.
2. SE’s moderation works much better than MB’s
SE sites are moderated by the community, perhaps not all of it (you have to have a certain number of reputation points to be able to moedrate a post), but usually the greatest part. SE allows for editing of posts made by others (saves the reader from posts in an MB thread complaining about typos for example), and most importantly allows for voting on questions, answers, and comments. Additionally, it includes more classical type of moderation, such as moving, or closing posts (eliminates duplication). In contrast to MB-style sites, however, a large part of the community can perform these kinds of moderation.
(users ∩ moderators)SE > (users ∩ moderators)MB*MB sites on the other hand are moderated by moderators. While the principle of moderation is very basically the same – moderators are part of the community – the intersection of users and moderators on MBs is much, much smaller than on SE-based sites. This allows for much faster, better, and more democratic moderation at SE sites. Argumentum e contrario this means that the mean quality of answers on SE sites will naturally be higher than that of MBs.
* Possibly highly incorrect, but I like the look of it.
3. SE answers retain a high visibility
When you google a question, and are given a link to an answer on an SE-based site, it will be to the single site that also gives you the answer (just below the question), no matter how old the question/answer is, and no matter how many other (less-quality) answers have been added after the “best”, “correct” answer. On MBs, there may have followed a number of other posts after the “best possible” answer has been given, and it wil be harder to find the best possible answer, because you may have to read through a 20-page thread to get there. So the chance that you can get from typing your question into a search engine site and to a good answer with one or two clicks is much higher for SE sites than for MBs.
I like MBs, I really do! They’re great for discussion of any kind, including techie stuff, but also job proposals, socialising, networking. I just think they aren’t that well-suited for pure Q&A, and that’s because of the noise! Even the most diligently moderated MB will contain n orders of magnitude more noise than any SE-based site, n depending on the number of moderators and their diligence. Because on SE sites, the greater part of the community will moderate the content – both by usual means of moderation such as editing, moving and closing questions and answers as well as by voting on all types of content – which means that any noise that appears will probably be taken good care of in relatively little time. There is just little to no room for off-topic discussion, opinion pieces, spam, and hubris (i.e., noise), when almost anyone will be able to simply edit that out. Provided that the community is large enough and well-formed. Which leads us to…
5. SE sites are prone to have stronger communities faster
This will be my last point before I turn into a drooling bootlicker in front of the whole interwebs. Because the community for any SE site must be strong enough before the site even goes live, and individuals in the community must have made a real effort to get it to that point – and may therefore arguably be more dedicated to the site’s well-being – the whole grassroots moderation concept of SE sites is likely to be succeeding when a site goes live. With MBs, much the same is obviously possible, but there it is more likely for an MB to stall and die for a lack of moderators or community discipline.
Proposal for a Stack Exchange site for the Eclipse Community’s Q&A
So, seeing that I like SE a lot, almost as much as I do like Eclipse and its great community, it would be a shame if the two couldn’t be brought together some way or another.
That’s why I’d like to propose a move of the Q&A section of the Eclipse Community Forums, currently running on FUDForum software, to a Stack Exchange site.I believe that the Eclipse community will gain massively in doing so, for the following reasons (again, not comprehensive).
- An Eclipse Stack Exchange site will be the one stop shop for all concrete questions related to Eclipse technologies (i.e., the Eclipse IDE in its various phenotypes as well as all Eclipse and Eclipse-related projects). Including those Eclipse-related projects which are not hosted at Eclipse will ensure that such a site will cover close to all things Eclipse.
- The Eclipse Community will be able to leverage the above-mentioned advantages an SE site has over an MB: No more browsing long threads, no more badly placed questions, no more duplicates, no more off-topic noise, easy-to-find answers that have been approved by the community (in a best case scenario). The only argument that’s void is the one listed under 5 above: The Eclipse community is already very strong, and I anticipate that they know their way around SE sites already, so building the community already got a tick on the list.
- The Eclipse Community Forums (ECF) will become more readable, as the long lists of forums will be transformed into a list of tags on the SE site. Therefore the ECF will remain the first place to go for non-Q&A-related interaction. Questions that aren’t strictly technical, such as project-related proposals, could be asked on a discussions site (similar to discuss.area51.stackexchange.com), and there would be a meta site, which would provide a Q&A site about the Q&A site (similar to meta.stackoverflow.com). As for the newsgroup (nntp) functionality provided by the ECF at the moment, I’d suggest every newsgroup gets a section on the project website it relates to.
I’d like to propose the following split of the existing ECF over the two sites.
|Eclipse Community Forums|
|Eclipse Working Groups||Eclipse Projects|
*Including the inactive(d) threads not to be moved to the Eclipse SE site
Possibilities for setup
There are two general possibilities for setting up an Eclipse SE site, with the first one being the more favourable by far, in my opinion:
- The Q&A section of the ECF will be moved to an “official” SE site under the eclipse.org domain, and therefore removed from the ECF. Active questions can be moved to the SE site, while questions that have been inactive for a certain amount of time will be archived, and made accessible from the ECF, possibly via the Archived forum. The Eclipse Foundation will manage both the SE site and the reduced ECF.
- There may be political or legal reasons why an Eclipse SE site under the aegis of the Eclipse Foundation is not possible, e.g., copyright of the site, SE not being open source, etc.
- There may be reasons why some of the needed functionality may not be possible to be made available by Stack Exchange, e.g.: a custom domain (stackexchange.eclipse.org for want of a better idea), adding the existing user base from the ECF to the SE site (provided that this is legally possible and wanted in the first place), etc.
- Users will have to get used to the new site, which is a neglectable disadvantage compared to the advantages they would gain.
- An “unofficial” Eclipse SE site will be set up in addition to the ECF, which will remain as it is.
- The two sites are competing, there is no one stop shop, and users may have to remember one more login to one more site, providing that the Eclipse SE site is embraced by the community, which I anticipate to happen.
- There may be too strong an overlap with Stack Overflow, which could prevent an Eclipse SE site from getting off the ground in general.
- (The ECF may lose traffic.) [Not sure if that's a viable point.]
The future of the Eclipse Community Forums
In the context of this proposal, the ECF will remain the main discussion/networking/socialising site for Eclipse. Only the Q&A content will move to the SE site. Not a lot will change really, apart from the existence of a better-suited platform for strict Q&A.
Is this a good idea?
I’d like to discuss this proposal with people that know a lot more about communities, the law, and software politics than myself, so please feel free to comment, either below in the comments section, or on the Eclipse Community Forums, in this topic: A Stack Exchange site for the Eclipse community?!.
I also thought it would be a good idea to ask nicely first, so I’ve sent this proposal to the Eclipse Webmaster before publishing this post. I’ve also sent a few preliminary questions to the folks at Area 51. If permitted, I’ll post their replies on this blog.
There is one major drawback about the proposal which, while pondering about it again before publishing this post, has come to my mind albeit being the most obvious one; There is already a SE covering all Eclipse questions: stackoverflow.com.
The problem is, of course, that for people who are after Eclipse-specific questions only, there is a lot (a LOT) of “noise” on this site, it being the premier one-stop shop for pretty much all programming-related questions. The existence of stackoverflow.com alone may prevent the setup of an Eclipse-specific SE site, as it is SE’s policy to not set up sites whose community is already catered for (cf. the Area 51 FAQs, especially the paragraph titled “Should my idea be part of an existing site, or its own site?”).
However, the positive effect an SE site solely dedicated to Eclipse-related questions will have on the Eclipse community does – IMHO – easily outweigh the negative effects the existence of such a site will have on stackoverflow.com, as well as the resulting (fairly minimal, cf. next paragraph) fragmentation of programmer communities over One More Site.
In total, there are 51,022 Eclipse-related questions on stackoverflow.com, based on the 50 tags including the phrase “eclipse” (cf. table below). This compares to 4.25m (4,250,000) questions spread over 31,950 tags relating to non-Eclipse-related topics (cf. graphs below). At the same time, there are at the time of writing a total of 206,598 topics in the Eclipse Community’s forums Language IDEs, Modeling, Eclipse Projects (i.e., those that I suggested to be moved to an Eclipse SE), spread over 169 topic forums.
|Eclipse-related tags on stackoverflow.com|
|eclipse (38083)||eclipse-plugin (4426)||eclipse-rcp (1806)||eclipselink (1397)|
|eclipse-cdt (966)||m2eclipse (653)||blackberry-eclipse-plugin (458)||eclipse-juno (290)|
|eclipse-indigo (289)||eclipse-wtp (259)||eclipse-emf (251)||eclipse-pdt (248)|
|myeclipse (188)||eclipse-pde (179)||eclipse-plugin-dev (171)||eclipse-gef (141)|
|eclipse-3.6 (135)||eclipse-jdt (122)||eclipse-emf-ecore (90)||eclipse-adt (86)|
|eclipse-3.4 (79)||eclipse-3.5 (75)||eclipse-virgo (56)||eclipse-gmf (55)|
|google-eclipse-plugin (54)||eclipse-rap (49)||maven-eclipse-plugin (37)||gae-eclipse-plugin (32)|
|mercurialeclipse (32)||eclipse-tptp (26)||google-plugin-eclipse (24)||eclipse-memory-analyzer (23)|
|phpeclipse (19)||eclipse-rse (19)||eclipse-marketplace (18)||cfeclipse (14)|
|eclipse-ecf (14)||eclipse-mat (13)||eclipseme (12)||eclipse-gemini (12)|
|sbteclipse (12)||eclipse-dtp (10)||eclipse-m2t-jet (9)||eclipse-templates (8)|
|hgeclipse (8)||ant4eclipse (8)||eclipse-3.3 (7)||eclipse-europa (7)|
|eclipse-fp (7)||eclipse-databinding (7)|
|Tags including the phrase “eclipse” on stackoverflow.com. Source: http://stackoverflow.com/tags. Accessed: 22 Jan 2013.|
Maybe the proposal of a Stack Exchange site for the Eclipse Community isn’t flawless. There are obviously drawbacks, and there are surely interests attached to it which might be opposed to such a site. In my opinion, however, all parties that would have to be involved in this, Stack Exchange and Stack Overflow, the Eclipse Foundation and the Eclipse community, will be able to leverage it: SE and SO by drawing more attention to their great software by getting involved with one of the best active developer communities out there, the Foundation by providing developers and users of their technology a great Q&A site, and the Eclipse community by gaining a one-stop shop for all their questions.
I hope for some discussion about it, either here or over at the ECF. Let me know what you think!
Update 1: SE is not OSS
Fabian Steeg has mentioned in a Twitter post what could be a major pitfall for a specific part of the whole idea: SE is not Open Source Software! This is something that might be keep the Eclipse Foundation from using SE for an official Q&A site (although there is nothing specific mentioned in the About section of the Foundation’s website at http://www.eclipse.org/org/#IT). My first reaction to this issue would be to think, “Well, let’s just make it unofficial then”. I guess if this strategy was pursued while at the same time the Eclipse Community Forums would close the forums mentioned above, a similar effect to the one I had in mind originally might be gained. However, the latter part of this of course remains wholly at the discretion of the Eclipse Foundation.
It is to be found out whether the Eclipse Foundation disallows use of non-OSS for its purposes as well. Comments about this, as well as any other aspect of the proposal, are very welcome!
Update 2: Bug
I’ve created a bug for this proposal at https://bugs.eclipse.org/bugs/show_bug.cgi?id=398767.
|If you found this post helpful, you may want to consider donating to help cover server costs.|