Talk:MediaWiki modifications

From Consumerium development wiki R&D Wiki
Jump to navigation Jump to search

This page is for discussing and defining the modifications/patches that we need to make to the MediaWiki software to enable running Content Wiki and Opinion Wiki. A similar page exists for each other wiki code option.

It's presently being debated. See Features and Consumerium Services for the more stable list that motivates these requests.

corpus coherence

The interwiki link standard (very minor change), interwiki identity standard and other features on list here which also lists other features like Simple English and Symbolwiki conventions that simplify the UI, very important for worn devices, and other things that can be done without code modification.

tracking all changes to the corpus

So you can see if for instance a Disinfopedia or Wikipedia or Recyclopedia article on a company or promoter changes, plus the several Consumerium Wikis.

See GFDL corpus watch for a feature proposal that GetWiki is more likely to do, and is absolutely required for any multiple-wiki project.

interwiki link standard

as per convention, use links in section titles only if the page MUST be read to understand what follows, if the features invoked are add-ons to the concept

Without this, your namespaces aren't coherent across Consumerium Services:


This is a new feature in MediaWiki. It is a namespace for storing and accessing large number of short items quickly. All [[MediaWiki:]] entries are stored in RAM and can be modified on the fly by sysops. MediaWiki-namespace items have the possibility to use variable-subsitution which is really useful for displaying campaign stats in Opinion Wiki articles among other things.

For further information on this: m:Meta-Wikimedia:MediaWiki namespace

Added namespaces

For Content Wiki to be managed well we need to have at least the following namespaces:

  • Product
  • ProductGroup for product groups
  • Company
  • faction and tendency (describing one's own biases/assumptions)

For the Opinion Wiki we need at least the following namespaces:

  • Group
  • VirtualGroup - virtual group is a group that is not a real world group, but usually an ad hoc action group to drive some campaign goals
  • Campaign

moving articles between wikis

standard wiki URI

Without this, you aren't sure that you're coordinating two articles about the same thing.

Content autogeneration

We need to have a module for generating and identifying information coming from the Consumerium Vault, wherein information is verified by staff. Also reacting to editing of autogenrated content (if made possible) has to be dealt with.

Middleware connection with Opinion Wiki edits

Edits to the Opinion Wiki have to be screened by middleware software to update the composite view of the wiki eg. changes in votes or scores

interwiki identity standard

Without this, you can't keep track of who is who in the various wikis, or do:

permission-based model

These features are controversial, and maybe already in tikiwiki:

Article signing

The software has to be extended to enable organisations and individuals to sign any version of any article in the Content Wiki with GnuPG or other keys. This is essential to maintain the integrity of the recorded information. Some people may want to see the live wiki, understanding the possible hazards in it, but most will likely want some level of assurance from authorative parties on the correctness of the information

Article creation restrictions

For Opinion Wiki we need to restrict the creation of Group-articles to only those entities that are registered in and verified by the Vault

Privilege management

For Opinion Wiki we need totally new code to manage the addition, modification and removal of usergroups who can be granted arbitrary priveleges (at present there is only 'developer', 'sysop' and 'blocked IP' status). Mediawiki stores both article privileges and user privileges as a comma-separated list, so it's not much to hack this

Editing restrictions

For the Opinion Wiki we need to manage restriction of editing of certain articles and subarticles to certain users belonging to designed usergroups OR alternatively require their editorial approval before an edit is visible.

Given that the features you want are mostly in tikiwiki, why not just use that? It would be easier to support a common wikitext standard in that more supported software, than try to make mediawiki do what it doesn't do well.

I'm looking into it. The new version 1.8 or Polaris seems much more responsive then the 1.4 and 1.5 i fooled around with when trying to locate a CMS to do this concept development & research stuff.
  • It does do valid XHTML which is an advantage over
Yes a huge edge. And it will likely do XForms long before other wikis, for this reason. XML integration is important. Microsoft Internet Explorer is so integrated with XML now, that we can just require that people use it if they want to work on opinion/content, and that will cut the coding very drastically.
Whaddefuck?You want to contirbants be depended on M$ Internet Esplorer. You are not sane you troll. I havent looked into what xforms actually is but from the name i guess it's nothing special that can't be done with cleaver HTML Forms. But hey whaddoiknow. I'll look into the issue.
XForms' main advantages are (a) multi-page forms that can be dug through without server interaction, submitting the whole thing back at once, and ending that "page expired" problem. (b) separating purpose of a form from style and (c) submitting the actual <instance> </instance> as XML. But IE doesn't do this yet. What IE does right now is work very well with XML, and a lot of people standardize on it only for that.
  • The linkage visualization (map) thing seems very interesting
  • It does have group management already, but according to my knowledge adding this to MediaWiki isn't a problem at all.
No, that's wrong. Look at the MediaWiki user base - mostly at Wikipedia. These people do group management very very badly and freezing their bad ideas about it into code will make it worse. For years they have very serious governance problems, there are always big troll fights and "regime change" debates and flame wars, and "pogroms" and "witchhunts" and "purges". Comments on "what's really wrong" get censored by a group that doesn't want to hear it.
They just believe the plurality of contributors will keep the project alive well.
So, they have no actual concept of how responsibilities break down, other than this:
the groups they have being:
  • Anonymous
  • User
  • Sysop
  • Developer
You forget, when anonymous offends sysop, it becomes troll. LOL! And there is no editor or lawyer which there might have to be soon.
I'll describe how nearly unlimited amount of Groups can be implemented in MediaWiki without any major modifications when I feel up to it. It's not really a problem IMHO.
They just don't know what they're doing, and on MeatballWiki and such you can find people complaining about how stupid the Wikipedia people are about how to do real world group management. They're always the worst example, e.g. of GodKing or just being a libel pit where anyone can lie about anyone else without any consequences. They'll collapse the day some guy with lawyers notices what they have allowed to be said about him. Like go look at the Page History of the article on Mel Gibson!!! And Mel sues, for real... he even sues CHURCHES...
People who organize their own favourite project so stupidly can't be trusted to figure out what "requirements" exist for serious social software! This is a very good reason to get away from MediaWiki and not to trust the people working on it. They are just not politically mature or even legally responsible. A project like Consumerium which is even more of a target can't possibly rely on software created by people who have such a poor idea of group support.
more on this debate MediWiki vs. Other Wiki Implementations later on. Maybe we should commit an article to this issue.Juxo
Yes, but keep in mind that much more than just "a wiki implementation" is required. There's a need for signed and authenticated articles probably, and, maybe even payments processing for donations, or (as I claim) for "betting".

MoinMoinWiki is even more extensible with a GREAT parser and extension system.

This really should be looked at. It's Python, but why not copy the exact architecture of it into PHP or Java, and integrate it into tikiwiki or etc?

Some of these "features" are disastrously wrong. While it may make sense to bar certain classes (like anonymous IPs) from making modifications that are immediately publicly visible, it never makes sense to simply reject what they wrote without reading it. Having "a group that is the only accepted list of authors" is just WRONG, not wiki way but also WRONG from the point of view of a system that is supposed to be outreaching all the time and including not-well-known views. Nothing prevents the system from simply not publishing some comment to the world until someone else approves it (say by just signing it).

Of course if you talk to the morons who choose misfeatures at Wikipedia, and who are destroying that project, you will get a wrong idea of what is "needed".

And tikiwiki has most of these misfeatures already.

There is an effort to make MoinMoin read a wikitext standard based on the version of MediaWiki that is used at Wikipedia, to rescue all the good articles from all those bad people. That would be immnsely useful, much more so than trusting them further to make extensive mods to mediawiki. Which doesn't even support full text search in their present configuration. It's very hard to imagine they'd ever put consumerium first on their list of feature needs.

OK, this entire page is nonsense. The features are needed, don't matter, will be low priority compared to Wikimedia requests, and already exist in other wiki code. The real features required are integration with mobile device code and ways to handle hordes of battling users (faction being only one proposal for this) with different POVs they aren't compromising on.

It would be sysopish to simply trash this page and correct all the errors; What sysops just did to bet is not a good tactic; So what's the right way?

The right way was to integrate it all, and note the concept (permission-based model) one is "buying into" by implementing certain misfeatures. Done.

A better list is at en: Recyclopedia: Talk: Recyclopedia: Feature. Suggest that be adopted here, and expanded for Consumerium needs. If neither MediaWiki or GetWiki wants to do it, we fork or shift to tikiwiki or MoinMoin.

Listed up front. Well worth reading. Should also be more fully integrated.