There are several things that are not so nice in MediaWiki, edit conflicts is one and wikitext is another. The MediaWiki Wave project is happy to show you that edit conflicts and several types of wikitext can be a thing of the past.
Wave provides a great colloborative editing interface, with a realtext interface. This is the MediaWiki that everybody can edit.. :)
Today we are to produce a little bit of video about our MediaWiki Wave product. Everything was working, we had two bits of code that were trivial to merge.. but it was not.
When code is alpha, things change and Wave is still very much alpha software. Tom is working hard to get rid of errors and to get the functionality to work again. The deadline for delivering our video is nigh.
So Murphy is not welcome.. Please Mr Murphy go away !!
One of the persistent noises at Wikimedia mailing lists are those made by those who see a conspiracy everywhere. Who always assume things that there is more to the story. One target that is increasingly considered evil is Google, even though one of Google's aims is not to be evil. And given their size, it is assumed that everything is targeted to bring about world domination.
The one “attempt at world domination” that I personally hope will succeed is Wave. I love it because it brings so many aspects together of applications that are on their own incomplete and/or dysfunctional like e-mail, forums, chat and wikis. For me, Wave provides a platform, a protocol that is intended to be extended. It allows you to add to it and make it better like we are doing with the MediaWiki Wave interface. This is possible because people are invited to add to the Wave protocol as long as they implement the functionality described on the public reference implementation as well. All the resulting software will be available under a free license and consequently even the FSF is interested in how Wave evolves.
World domination requires an evil genius but the genius behind Wave are Lars and Jens Rasmussen. They had enough of being the big guys behind Google Maps and decided to do something new. Their thing was to do technical things and not grow pointy hair. They called their project Walkabout... Walkabout is this Australian thing where you remove yourself from civilisation to experience the world to come back and tell about it.
They certainly told the world about it. They renamed Walkabout Wave and they mixed and merged many of the paradigms of computer communication; e-mail, chat and used that as a starting point because they also brought Wiki and collaborative editing in the mix.
Wave is heady stuff. I find I am still grappling with its potential because while on the one hand it provides a much richer experience then e-mail or chat, the interface on the other hand suggests that it is still centred on small groups of people sharing the environment. In most environments, business or personal that makes sense.
We are however integrating Wave with MediaWiki. Wikipedia articles can have hundreds of contributors and they do not really know each other. In a wiki all articles are mine to edit, the authorisation is not based on me as an individual but on me as part of a group. What we already demonstrate in our project are the immense benefits Wave provides. As we progress it becomes clear that the inclusion of a Wiki in an environment that started of with the merger of the electronic equivalents of a letter and a telephone conversation is similar to throwing a newspaper into the mix.
We have to get this right because how else are we going to conquer the world
When you demo, you show a scenario. You know how to wow the public and you consider what is of interest and makes people anticipate the moment when they can get their hands on the code. The moment of truth, when the software is not used within a scenario, where the article that is edited is no longer called "elephant", but can be any article.
Consequently, it is no longer convenient to hard code "elephant" because people might want to write an article called "tsunami" or "spring tide" when they think that "wave" is a bit tame. There are five such hard coded gotchas. In stead of a predetermined solution, we want the same flexibility as in all Wikis; you determine the name of the article.
So our demoes are rigged, but we are removing the rigging and make it what you see is what you get.
So first an update, we have code that does syntax highlighting. Templates are no longer a real problem and neither are citations. What is a problem is to make a button that allows people to switch from WYSIWYG to Wiki syntax and vice versa. It is quite crucial that we get over this hump because it prevents us from rolling out new functionality.
In addition to the 50 testers from Combat testing, we have 9 Wikimedians willing to test. Nie who either asked for a profile at Wikimania or asked for a Wave account. Only nine of the original fifteen people registered at our forum as asked. We have indications that all 15 now have a Wave account. So when you did not do so, please register at the Forum.
The testing plan will consist of two phases:
get baseline measurements on the usability of the first iteration of the WYSIWYG functionality and learn/ask what should be / could be different
get measurements on the usability once the lessons learned are incorporated in the WYSIWYG functionality
We hope / expect that by Monday we will have something to show for our efforts.
When Tom learned that Douwe Osinga was in the Netherlands, it was a great opportunity to seek a meeting and learn what the time line is for Wave.
If one thing is clear, it is that by the end of September 100.000 people will gain access to the Wave environment. Beyond that a lot of hard work is still to be done to make Wave ready. For us in the MediaWiki Wave project, it is clear that we will not wait for the APIs that will be ready in "two weeks". What we can do is sanitise the existing code, and include new functionality like the syntax highlighter written by Kim. There are plenty of things we can add to our code, using MediaWiki authentication is one of them and we will.
When you meet people like Douwe, it is a great moment to ask and inform about the issues on our end as well. There are two key show stoppers that will prevent adoption of Wave in the Wikimedia Foundation they are:
Wave must be published open source
Wave must include internationalisation infrastructure
In the way Google is developing Wave, they want to provide a stable product that will provide people a stable basis for further development. Some of the software provides this basis and this allowed for our project while other parts are just not ready. This is a reasonable rationale but it is equaly understandable that the WMF will not adopt software that is not published open source.
Given that there are Wikipedias in over 250 languages, all software has to be internationalised. At this time, the core Wave functionality includes the message structures that provide the basic building blocks for internationalisation. It is not clear how these building blocks are to be used in our software. It is not yet clear how we will interface with the Wave data; in MediaWiki we distinguish between the language of the content and the language of the interface. Wave does allow for multiple languages in the content.
What will be interesting is how we will resolve the issues we have with projects that use incorrect language codes. This is an issue that the WMF should address anyway. The easiest option would be to support only those projects with advanced software that conform to the relevant standards. Thanks, GerardM