Gossamer Forum
Home : Products : Gossamer Links : Development, Plugins and Globals :

Blogs, Blogs, Blogs, what about Wikis??

Quote Reply
Blogs, Blogs, Blogs, what about Wikis??
Everyone is looking at "blogs" and I wonder if it's the same myopia that causes people to obsess about SEO, and forget the goal is to create content-rich sites that users want to keep visiting and/or participate in.

Blogs are basically CMS (content management systems) that have a few extra features (and a whole host of missing ones). Specifically, the trackback, ping, blogroll, and other "announce" features are what most people fixate on, et, without something to announce, they don't mean much. Oddly, except for a few special instances, such as a breaking-news blog, those features really don't do much for the blog itself, and really increase the noise to signal ratio, and make it more confusing for a novice surfer (90% of what is out there) and very often the real target of the blog. Being indexed by the search engines for your content is how most people are going end up at your blog.

Most blogs are personal logs. Blogger software has been upgraded to do other things, but they were, and are, really a way for someone to get their own soap box, or newsletter, and get some feedback on it beyond the simple webpage.

What's the point of "publishing" your blog, if it's just being thrown in with a million other blogs, sort of like standing in the middle of the down town and tossing your entire handfull of leaflets in the air, and walking away. There is a much higher yeild if you hand them out one by one.

Right now, Links can do most of this, with the reviews system, and the soon to be released comments system, there is very little that can't be done within the core links modules to create an interactive blog-like system (without the "publish" features which I'm really still trying to understand their function).

Ok, that said, what *most* requests seem to be, are *really* Wikis. Where people can interact with, and comment on, and update/change the base content.

For instance, the http://www.wikipedia.org is the classic example, and everyone has bumped into it at one time or another.

While a wiki is too close to anarchy for me, there is another type of system, exemplified by Scoop, that is between a closed CMS and an open "wiki". What it allows is for the user community to post, vote on, publish, remove, etc articles in a somewhat majority/democracy rules system. Some people talk about this as a completely admin-less system, where it's fully user-driven. It's a very advanced "karma" system, where users, messages, posts, comments and such are given "karma" like points, and that determines if they live or die, where they are ranked, and if they come back to life.

What people seem to be asking for, rather than the spam-announce-publish system of the blogs, or the total anarchy of the true wiki, is something along the lines of Scoop. http://scoop.kuro5hin.org/ It seems to be a CCMS -- collaborative content management system.

A description/explanation is here: http://scoop.kuro5hin.org/special/whatisit

The IMDB (http://imdb.com) may be the biggest of the projects, but uses a fairly closed system, where your updates are reviewed by a panel, not the user base at large.

I think the current implementation of scoop is too cumbersome to install, and too locked-down in it's legacy code. A version 2.0 seems to be being planned, but looks like it will be a complete rewrite.

In analyzing what is going on with the various Wikis and with Scoop, it seems possible to create what a version 2.0 of Scoop should be with Links. *AND* the install, maint, and upgrades would be easier.


Most wikis are not using a MySQL database, and that limits their size and power. It's the same issues that started to limit Links 2.0 before Links SQL was introduced.

The idea is to look at a Wiki in terms of Links SQL, and controlling the "anarchy" by using "karma" points.

Because of some limitations in Links, it would be helpful to do a little mapping of terminology.
  • Wikis have "webs" which are similar to what we'd consider "Main Topics" or root level catagories.
  • Wikis talk about "topics" which appear to be just web pages, but there are subtle differences.
  • Topics are linked by using "WikiWords" which are strings of capitalized words run together ThisWouldBeANewTopic
  • In display, ThisWouldBeANewTopic would be hyperlinked to the "view" script (page.cgi) with ThisWouldBeANewTopic as the p= parameter.

This is where it starts to break down, if the use of a MySQL database was to be optimized.



I'm going to stop here, before wandering into more technical details, in order to get some feedback. Blogs, Wikis, CMS, etc. What is it you are looking for in your site development for how users interact with you, and vice versa.


PUGDOG� Enterprises, Inc.

The best way to contact me is to NOT use Email.
Please leave a PM here.
Quote Reply
Re: [pugdog] Blogs, Blogs, Blogs, what about Wikis?? In reply to
Hi pugdog,

I join you in your rant about blogs which seems to be thriving on more of a fashion type hype than any real technological innovation. I worked with a system called StoryServer from Vignette that is CMS type system but when I was started on it, it lacked a lot of workflow capacity that was a goldmine (while this was lacking) for web agencies et al. since it was difficult to imagine using it without a workflow system for newspapers. Guardian unlimited (UK) Le Monde (FR) and the Wall Street Journal (US) use this at the moment. Although I don’t think GT is holding back on key functions deliberately it would be nice to see a) group management, on top of existing user management available and b) workflow capacity which would be a great boost for the product and provide an endless amount of new function ideas for plug-in developers. I do feel that these are core functions. Seeing them developed in a sound and secure way by GT would be great and would then provide developers from GT and the plug-in community with a great basis to work from.
I’m not sure how your (future?) Blog system works, as I told you I had started working on specifications with GT prior to your announcement for a Blog system that only required a few mods since as you said a lot of blog functionality is already present in Links compared to present blog systems on the market. In my specs I had knocked out requirements for trackback and ping which friends of my that have blogs say they and their readers never use...
I think that the fact companies can envisage creating a blog system so tightly linked to the product is in an indication of how impressive Links and the code it’s built on really are…
Significant Media
Quote Reply
Re: [pugdog] Blogs, Blogs, Blogs, what about Wikis?? In reply to
pugdog, we think too much alike, lol.

I am looking at needing a way to get the wikipedia data imported into links the same way you import dmoz. I would love to be able to import sections and then let my directory/community members make changes. There is a page at wikipedia that has the slices and explains how to import parts, but I haven't gotten that far yet.

There are smaller bits you can start with, like the wikiquotes... It's only 85 megs...
Quote Reply
Re: [loxly] Blogs, Blogs, Blogs, what about Wikis?? In reply to
maybe andy can help on that. He's becoming the master on imports and exports :)

I'm seriously looking at "forking" a wiki and Scoop into Links SQL.

The program is more like a "blinki" -- blog-link-wiki program.

Code:

l | Blog | L W | L | W | W |
wiki | i | BLINKI | B I W | B I | B I | BLOG
n | n | L N K | BLINKI | LINKS | LINK | I
k | wiki | O K I | O K K | O I | O I | N
s | s | G S | G S I | G | G | WIKI


As shown by the prospective logo Skelletons, it's a combination of a Blog/Blogger, Links and a Wiki.
I've registered blinki.info for the discussion of this project, but will probably call it something
like the "BlinkiBlogger" or "BlinkiWiki".

Or, actually its' more complex or convoluted, rather than a simple crossword, more of a word-find:

Code:
B I B BLINKI BLOGGER
L K L L L
I I O O I
W N N G G N
K G G K
S I E E I
R


I will take a look at the wikipedia dump, maybe get some ideas for how to set up database. There *are* a lot of ideas out there, but none take into account the underlying "Links" layout.

I've installed a few test blogs and wikis, both perl and php, and don't really like any of them.

I do need to find a karma system, for rating/ranking and ordering the wiki. There was a discussion about this awhile back, what constitutes a "karma" system.

It has to take a load of things into account.

1) how active a user is on the system
2) how relavent and well-liked the user is on the system eg: how others see him
3) how relevant a post is considered, weighted by the ratings of the user casting the vote
4) etc.

Thresholds have to be set up that allow certain users to do things that others can't, and which allow posts and comments to float higher, or lower, or decide if an article is published, hidden, or totally trashed.

A users karma will depend on all these things. This in turn, would affect how "powerful" their ratings are.

There needs to be a way to set the parameters between the highest and lowest "karma" fudge factor, eg: 1-5, 0.1-1.0, etc. to set up how "powerful" the most karmic user is compared to the lowest. The number should never go above a certain ratio, or things get out of wack real fast. There are endless discussions about this on-line, but some simple, effective system beyond a simple +/- (since every user's ranking counts for as much).


PUGDOG� Enterprises, Inc.

The best way to contact me is to NOT use Email.
Please leave a PM here.
Quote Reply
Re: [pugdog] Blogs, Blogs, Blogs, what about Wikis?? In reply to
Just an update since it's the beginning of a new month.

I see Philip(fuzzy_logic) has updated his plans as well ;) But for those wondering, my project was never to compete with his, in fact, he revived his at just about the same time I got an interest in this through registering a domain name. We talked, he wanted to develop on his own. Competition/choices are good. I have no problems with that.

But, in the past month, I've done a lot of surfing, reading, and playing with Wiki's as opposed to 'blogs' (and not a lot of sleeping). My goal, for all this, was much more wiki-like than blog-like. I don't need/want the spammish announce type features, but I wanted to be able to allow a user to post/set up a news or diary page, and allow interactive comments on them. With a wiki, the ability to allow free-form editing is also possible.

I have set up a couple of blogs, and quite frankly, don't like them at all. It's not "me". I like to code things I like. Something about the blog seems stunted.

I've set up 4 or 5 different wikis, and while I don't like any of them, I *do* like the concept. They are collaborative, which translates to interactive, which translates into "stickiness" for surfers.

What don't I like?

First, almost all development of SQL backed wiki's is in PHP. The "H" stands for "HOLES". I don't like code in the web tree.
Second, PERL wikis use too many system hacks and do most things in a flat-file format
Third, security on them is not well thought out, and relies on .htaccess and RCS revisions roll back.
Forth, they all have their own rendering, parsing, and access utilities.

So, my goal, is to take the concept, and use Links/GT code to do most of the work of managing users and access, parsing the templates, and accessing the database.

How I do that has been holding me back.

I don't want to reinvent the wheel, so I have to take the GNU/GPL rendering (syntax parsing code) and put it into a stand-alone module that can be accessed from another program. This is *not* my strong point, and anyone who would like to take on that job would be rewarded handsomely in many ways. This is the code that takes the WikiStyleWords and turns them into linkable objects, and deals with the [[]] formatting, etc. There is really no reason to write another engine, just condense and modularize the existing ones as a stand-alone module. Any "fixes" to the original modules could be easily applied, and under the GNU/GPL terms, this is allowed.

Ignoring that issue for awhile, the larger concept of how to work this into Links has been troublesome.

I don't want to start off with an OEM version of Links, eg: Blinki-Wiki or Blinki-Blogger that requires a whole license, but rather as a plugin that simply allows the users to set up their own system as described in the "blogger" posts above.

The difference is, that the spam features won't even be considered, and there will be an option for interactive edits of the comments and pages, not just simple post.

The issue then becomes do I allow all these mini-wikis to interact with each other, or just be limited to their own data.

Because I'm paranoid, and like security, they won't be free-for-all.

Users will have to register to edit.
Edits can be moderated (eg: approved before applied)
Edits can be rolled back to a point (maybe 7 generations, and/or days whichever is LONGER)
Users can be blocked from editing the original comments, or others comments, etc. (eg: a normal, non-wiki system)
Users can be allowed to edit the main topic (original comment) but not the comments posted by others. (I sort of like that)

It really starts to get hairy here, since it begins to look more like a bunch of porcupines tossed into a small box, than a top-down, tree-like, or even web-like system.

The comments plugin will be released, soon! I'm fixing up the pre 3.0.1 template issues, and cleaning up some calls. This "Blinki" idea will start to extend from this plugin, but the plugin will also be left "as is" with improvements for those who just want comments.

If someone would like to work on the syntax parser, I'd really appreciate it, as would a lot of others I'm sure. Not just links users. Since this is all GNU/GPL code, it can be released back to CPAN with attributions, and be a useful module/addition to the tools there. Reinventing this is pointless, and it's a potentially really useful tool for everyone, so if there are no takers here, I'll have to try to find a kid this summer who needs a project. Basically, the "content" block of the displayed page, what would be contained in the current "Description" field, would have to be parsed for tags. A block of raw text in, a block of parsed/replaced text out. Many good brains have worked out all the regexes (well, most of them <G>), it's just a matter of collecting them, and applying them in a reusable modular form, and releasing them back to the community at large.


PUGDOG� Enterprises, Inc.

The best way to contact me is to NOT use Email.
Please leave a PM here.
Quote Reply
Re: [pugdog] Blogs, Blogs, Blogs, what about Wikis?? In reply to
did you ever get this project completed?