Aloha! Here is something I have been thinking about for quite some time...
Since this has become a candidate for TLDR here is the short version: KHC is no goody, I want to make it more goody, what we need is better looks and integration of web content. I would like to propose that as GSoC project. Long version: - Intro Maybe it is just me, but KHelpCenter does not really live up to its potential. It does not look all that visually compelling, I did not see the search feature working *properly* in years (though that might also be because someone at Kubuntu decided it would be a good idea to patch away the search tab ;)) and it does not bring all available information to the user. - About the Internet Let me in particular reason a bit on that last point as I find the current state of affairs rather interesting. At least to me it would appear that using online documentation rather than docbook documentation (or a combination of both) is become a lot more attractive these days. They certainly both have advantages and disadvantages, but what we should acknowledge is that the existence of services such as userbase.kde.org, or stackexchange.com if you will, adds a whole new dimension to how documentation (read: knowledge) gets shared. Whatever you may think of the whole online vs. offline documentation topic, there is a very valid gain for the user in bringing the two closer together. Someone who has internet access and knows how to use a search site will more likely try to look for information online rather than offline. This is however not mutually exclusive with having offline documentation nor with having an application dedicated to present relevant information to the user. Actually I would argue that there should be a symbiosis of them. - Mixin Offline documentation provides a solid base of information on a vast amount of topics/applications. But imagine how cool it would be if we could enrich this solid base with additional unicorn sparkles and rainbows from the mighty and all-knowing internet. For example: you are reading up on Amarok in KHC because for some reason the audio quality is not quite what it should be. You are skimming through the handbook but there is nothing regarding your problem. Oh! But what is this? A tiny box next to the handbook that goes "Common Problems with Amarok: Sound is crap[link] My N900 is not detected[link]". Cool. So you click the first link and get all the information you need to fix the problem. (Turns out someone was using Phonon's Xine backend :P). Interestingly enough you also find out that there is a brand new awesome version of Amarok available, you get directions on how to upgrade and live happily ever after. - Let's crowdsource it! We all know that content on the internet is not always all that helpful and not always all that well written, so we could/should do something about that to. How about letting people rate content, add the information on what they were searching for into the mix and here comes semantic help :) - And Video! Now, I have only been talking about textual information, but let us not forget about other types of media. In particular screencasts. I happen to know some people who actually tend to go to youtube to find out how to do something rather than google for it. And indeed videos are a very convenient way to share knowledge. It is a bit of shame that this is not more tightly integrated into the help center, is it not? - Summary So, let me sum this up. * I do not think KHC is all that awesome right now. * Dragging information in from the internet is a great opportunity * Supporting multi media content would be terrific * Adding semantics to information would also be jolly awesome. - Technical Thoughts Finally to get a bit of a technical blah in here. On the information "aggregation" side KHC should be very scalable. Getting information from userbase.kde.org and other KDE specific resources is good. But we should also keep in mind that most distributions have documentation of their own (online that is), or screencast services (I know Ubuntu got one). Those should also nicely integrate. A plugin based approach would seem very useful here. Have a couple of resource providers that a) get information b) search information (c) display information, I am not quite sure about this. Should there actually be formats that can not be broken down to a HTML representation it probably would make sense to allow, but not enforce, that a plugin can implement their own view for displaying the information... something to think about) Offline documentation search should not use htdig but maybe something like clucene for deployment reasons mostly. Mixing in online search results as provided by the resource provider plugins. What would be interesting is redoing the UI using QML (because it is so much bling :)), I am not quite sure if that would make all that much sense at this point though. What do you think about this? Would this be a good project for GSoC? I would love to get some feedback :) regards, Harald
signature.asc
Description: This is a digitally signed message part.
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<