Ask HN: Why do we not have a Wiki yet? | Hacker News Reader

Ask HN: Why do we not have a Wiki yet?

51 pointsjdc29 comments

The format of this site is great for news, but I've been thinking: this community could really compile some very high quality information related to technology and business.

If I built it, would you come?

View on HN

One of the problems here is that every week, someone creates a new Wiki, or Excel of HNers, or archive of posts, and a week later no one remembers it. What's going to stop the wiki from being forgotten?

I'm trying to fix this "Meta" problem with a webapp I created, which is basically a (votable) list of resources for HN or any other site. You can find it here: http://resourcey.com/site_details/2/news.ycombinator.com/. It's helping me solve this issue for myself, so at least I'm scratching my own itch :)

This is a start, but it suffers from the same problem that all of the wikis, spreadsheets, archives, and alternative interfaces do -- how do people find it in the first place?

But _this_ is the problem I see the Hacker News wiki should solve. Maintaining a wiki about all the subjects that get posted here would quickly lead to madness. However, maintaining a wiki related to all the meta information the site generates (such as pointing to previous attempts to hide the points associated with comments) would be extremely valuable.

However, the only way this problem gets truly solved is if we can get pg to add a link to a wiki or something from the site.

Well, it is possible that pg will add a link to Resourcey.

But my hope is this becomes a semi-known place to go for "resource links for hackers". That way you only have to remember one site, but whatever other site you go to, you can find interesting spin-off resources.

I think that this is a great idea, this is completly inline with HN. I added all the services I know.

A wiki just doesn't seem aligned with HN to me. It's a place to post links, not generate content. What are you trying to achieve?

I just do the following: http://www.google.com/#q=site%3Anews.ycombinator.com+topic

I guess if posts could be tagged you could have a nice browsing function. wiki doesn't seem like the right tool though, maybe a del.icio.us clone that monitors HN and auto-tags?

It _was_ originally just a place to post links, but I think it's clear by now that it's grown into something much more. Many of us value HN for the quality of the conversations -- there are so many other places (like our personal Twitter feeds) where we can find links of interest.

Much of the value we get is from the content we all generate through these comments. It's clear to me that there's room for improvement when it comes to giving structure to all the great stuff that's discussed here. My gut feeling is that a wiki isn't the right way to go about it either, but there is certainly a lot of quality content being generated here.

I think we've established it should be user-editable. What structure would you suggest?

What about a wiki for hot topics. Someone gave P!=NP as an example. Collecting everything we find on these subjects with some blurbs about them.

I know non-coders might downvote me for this, but github's recently open-sourced Gollum engine might be a good basis.

  http://github.com/github/gollum

Contributors could write their articles independently, send a pull request, and have their writing merged into the central wiki.

There's also a (minimal) web UI for easy edits, or for users who aren't comfortable with git.

Gollum seems like a solid choice to me even if we most people don't use the git-ness at all.

It allows lots of markup formats (though we probably would want to pick one so that we don't have one page in textile, another in markdown, and yet another in AsciiDoc, confusing people that try to contribute), does syntax highlighting (which will be extremely useful), allows posting TeX in case we make a foray into a group P!=NP proof, and does allow power users to edit the files in their text editor and git push the changes.

P.S. clickable link: http://github.com/github/gollum

And you can see the editor here: http://github.com/github/gollum/wiki/Home/_edit

I shudder at the thought of how the "Do I need a co-founder?" wiki would look. Absolute mayhem.

It could be organized into different points of views. Each section could articulate pros and cons, for example. It's all in how you organize it; don't organize it like a fact-based wikipedia article.

Well, what need of yours is not being met? Let's start with the problem you're having, and build something that can fix it. I'll even help you with the coding. :)

The need that I think we're identifying is that often you need to find good info about, say, the ins and outs of raising an angel round. There's lots of great info on HN but it's scattered about among many threads. So you can't just go to wiki.ycombinator.com/angel-round and find out everything you need to know. You have to search, sift through the results ignoring false-positives (say, comments that reference raising an angel round but don't tell you how to raise one), etc.

So I guess the simplest way to state the need is that we need an easy way to access the wealth of high-quality information that's on HN. One solution is a wiki that structures a (hopefully high-quality) subset of that information.

The best solution, were it feasible, would probably be a super-"smart" search engine where you could query "How do I raise an angel round?" and it would understand the query and scan HN's comments and submitted articles (since I'm sure people have submitted articles that give tips for raising an angel round) and return a lot of information that it has determined applies, is high-quality, and covers all of the issues thoroughly. But if you can write that, you should be applying it to the entire web and making billions, not just using it to search HN ;)

So that's an example of a solution that would be better. Which means it's clearly worth spending some time considering whether there are feasible solutions that are better than a wiki.

The main problem I see here is that, even if a wiki wasn't too nebulous a solution to this problem, it is almost certainly perceived as one. Do you want to keep up Yet Another Wiki(tm)? Because I sure don't.

But besides that, it actually probably is too nebulous a solution. Wikis are great for well-defined, discrete quantities, e.g., Princess Diana. You can define it, tell what happened, etc. How do you do the same for an article like "Raising Angel Investors"? You don't, and hence people write articles about it instead.

In response to a comment by the OP [see below, or http://news.ycombinator.com/item?id=1618744], I proposed that we simply start by addressing the problem in the most basic terms we can, i.e., an HN faq or thereabouts, much like Reddit's Programming faq [http://code.reddit.com/wiki/help/faqs/programming], but I'm very open to suggestions, and -- as an enthusiastic denizen of this community -- I'm also completely willing to help with the legwork required to make a plan that doesn't suck come to fruition.

Perhaps just adding a search box to the HN header or footer would help? It would certainly be better than clicking through to Google, then clicking to enter some search terms, then browsing a list of results on a completely different site.

The problem really is that googling with site:news.ycombinator.com or using Search YC isn't a very good approach.

Using my raising an angel round example again, say you search "raise angel round" http://searchyc.com/raise%20angel%20round. The results suck. There's an Ask HN that could have been good except that it has no responses. There's an article that sounds potentially useful, except that the link no longer works. And then there's also a lot of things like "Tweetdeck raises an angel round." "Posterous raises an angel round." Totally useless in this case.

We have good information, but not good structure. Structure encourages high-quality contribution and eases querying of the information.

I admire your initiative, but "structure" -- at least as you describe it here -- is not a problem. It's a solution to the problem of organizing content. Don't you think?

The question is, are Wikis the best answer to this problem? They're really very good at helping us categorize things that are very differentiable, and very definable. For example: The Hundred Years War make a good article in a wiki. "How to raise angel investors" does not. Instead, people write articles about it, and you take what you want out of them.

Why don't we start with an HN faq, much like the Reddit programming faq [http://code.reddit.com/wiki/help/faqs/programming], and if that's successful, we can build from there. My philosophy is: address specifically the problem that you (or someone) has (i.e., "I have questions, and don't know where to begin!"), because spreading out too much or being too ambitious can literally kill a project. Don't take my word for it: the svn guys talk about how those sorts of projects were the "worst" parts of SourceForge: [http://code.google.com/events/io/2009/sessions/MythGeniusPro... <-- somewhere in there]

So let's start with a specific need and address that. Once again, I'd be happy to help you if you elect to go forward, but we need to attack one thing at a time.

I actually have no attachment to structuring this information as a wiki, but I know of no better readily available, collaborative information sharing software.

Problems of structure (front or backend) and organizing the content are the inseparable in my mind.

Gabriel's archive is a great starting point I think. Lot's of labour has gone into that already, use it as your base.

http://www.gabrielweinberg.com/startupswiki/Ask_YC_Archive

Take the initiative and people will support you.

I'm working on something related - a repository for "Great Articles". It's currently in early ante-pre-alpha, and sitting partly idle while I use it for real and assess its shortcomings.

If you want to have a look at what I have, email me.

If its a wiki in the traditional sense, it will fail here. Why? Because its just another freaking wiki that needs updating.

If there was an HN-endorsed crawler that could aggregate posts/comments about particular subjects intelligently and allowed posts on that wiki to inject into the relevant parts of HN (say, to a comment or a post), then it could be a Wiki-style meta-filter on HN.

Just an idea. Otherwise a technical wiki would have nothing in common with HN other than audience and launch base. People won't go back there and it won't beat Wikipedia at being a wiki.

http://remembersaurus.com/askhn.html

http://www.gabrielweinberg.com/startupswiki/Ask_YC_Archive

Are you saying "we already have wikis and they're good enough" or sharing a starting point?

When I look at those pages I see a fracation of what an HN wiki could be.

I suggested something similar also - Resources for Hackers:

http://news.ycombinator.com/item?id=1256069

I suggested a wiki here too. pg's suggestion was to have a section for evergreen topics. Didn't get much attention at the time though.

Yes, definitely, because is hard to look after all great content and "save" it for later reading and use.

Answer HN: Simplicity.

We should have our own jargon file too :)

Reply on news.ycombinator.com