I appreciate this group is here to dicuss the actual visual design of KDE, but it got me thinking. What use is a great icon set or theme if it sucks to actually apply it?

The Situation

Currently there are two main ways you can get your icons / themes. Either by the built in kgethotnewstuff browser in the system settings, or by kde-look.org directly. (Ok, some of you might browse deviant art etc. But i'm going on the basis that if you type 'kde icons' into google you get kde-look.org)

The Problem

Kgethotnewstuff is OK. It's functional, but not exceptional. The UI is in need of a refresh I think, and we could do with adding the ability to filter by colour / style / tags? etc. Leaving the UI issues to one side at the moment, I think the major usability problem with Kgethotnewstuff is that some of the themes and icon sets on there are uninstallable through Kgethotnewstuff, which should not be the case at all. As an average user you should not have to go hunting for a .tar.gz and work out how to install it, or, in the case of those that don't seem to want to install, start playing around with the '.hidden' folders. It's disappointing more than anything, especially if you really like the theme.

Which leads me onto the next problem...

Kde-look.org - The fuel to Kgethotnewstuff's fire. Well what to say about this beast of a website? It seems to become the defacto place for people to place their themes on, but I reckon only because of it's prominence in google. I can't imagine people actually like putting their themes on the site, but where else can they put them where a large amount of people are going to see them? The site is riddled with spam, in fact I'm looking at a piece of spam on the front page of the website that reads '12 yo lolitas hardcore WATCH HOT BOOTY *** AT', which is just appalling. Another problem is that a lot of the themes and icon sets there are old, and haven't been updated in a longgg time. True, some might still work, and look ok, but some wont.

The solution?

Well the solution I'm thinking of is tricky, and it's not going to happen overnight. If visual design is going to gain prominence in KDE, then I think it really needs a permanent home in KDE to give it legitimacy. Gnome has got it's extensions.gnome.org, so maybe we should have our own look.kde.org? Start from a clean slate, and perhaps incorporate a review system into every submission to check that the theme is in the very least installable through Kgethotnewstuff, has the right tags etc. etc. before it is approved and goes live on the site. Or maybe if it is not installable for whatever reason, but otherwise complete it could go on a special page on the site and not get feed into Kgethotnewstuff?

I think a lot themes are going to break on the upgrade to Frameworks 5 , so I would say this is an ideal time to do this kind of thing.

Well, I've got that off my chest now. I've been frustrated with Kgethotnewstuff for a while as you can probably tell.

Yes the GHNS situation is not very good. I'd very much like having a repository of things which are actually installable through the integrated UI, and -- as you said -- maybe have some kind of moderation (in purely technical terms, i.e. the item must at least be installable and usable -- no actual review process). I'd be surprised if a lot of KDE people were happy with the current situation.

The problem is that this is a lot of work, both on the technical and on the organizational side, and somebody needs to do it. So, the idea itself is honorable, but for anything to actually change, somebody needs to step up and organize something

Greetings,
Sven

Organizational can be fixed I think. It will have to be a long going process but when things on our end calm down a tad it wouldn't be completely insane if tried to handle those bits. Technically thought... phew! That is waaaay out my personal league.

But I completely agree - beyond looking more sternly at how the System Settings should be set up, how theming is handled is another HUGE thing. For most new users it is simply too complex and the feeling is that its not as theme'able as Xfce which is not true of course but since Xfce is easier to fiddle with thats the sensation.

One thought that occurred to me is the idea of Open Curation:

- Once every KDE Workspace release (~every 6 months) this group nominates a set of themes from any category(styles, desktop themes, window decs, etc.) to feature/promote/award/tag
- This group votes over a set period on 3 to 5 themes worth highlighting.
- Winners are featured as part of the Workspace release announcement to get maximum exposure. We could also maybe announce them on the VDG blog.
- GHNS gains a new facility to show VDG curated themes (maybe for just that release of the Workspace)

It might give an opportunity for some of folks in the wider community to get a nice spotlight for their work and provide a regular cycle where such contributions can be recognized beyond just the increasing less useful star rating and number of downloads metric.

The problem with GHNS is that it's heavily based around OCS (http://www.freedesktop.org/wiki/Specifi ... vices-1.7/). OCS is, in my opinion, not a great specification. When you implement it, you will end up with something very similar to kde-look.org -- not terrible, but not that awesome either. Additionally, collaboration with kde-look.org is difficult because it's closed-source and run by a single person, not the KDE community or sysadmins (still, you could try asking of course). So, to add some official KDE collection of themes which is actually controlled by KDE to the current GHNS widget, you need to set up a website which supports OCS (and there's not really a well-working piece of software out there which does this for you, so I would expect you will need to write a significant amount of code). The alternative is of course to use an alternate specification (which I would consider more sustainable long-term, because OCS is, as I said, not very good in my opinion), but this is might be even more technical work, since you need to implement the server part *and* a new client library for GHNS. On the other hand, if you don't want to support uploads (only downloads of administrator-defined content) and don't need a website (i.e. only access is through the widget) it might be doable.
I don't want to stop your enthusiasm, but this is the situation how I see it and it means all ideas which hope to actually get implemented at some point probably need a person in mind who is actually willing and capable of doing the work.

Another thing which was mentioned in this thread already is that there's no way to e.g. couple a widget style configuration to a colour scheme. This certainly could be improved with much less risk of failure.

Greetings,
Sven

I just rememebered that one of the devs on google+ was working on a framework for delivering content called Bodega. Here's a blog post about it here:

http://aseigo.blogspot.co.uk/2013/10/bo ... pants.html

Perhaps this would be a good place to start?

Yes I was also thinking about that. You'd have to ask Aaron about whether he thinks it'll make it to the desktop anytime soon. I think remember him talking about replacing OCS with this-ish, but what has been done until now seems to be pretty much geared towards "write an app store for a tablet". I might be wrong though.

Greetings,
Sven

Well it's a place to start though right? In the long run something has to be done, I think we all agree on that, so if we can at least gather some ideas here it gives us a reference point for the future. Perhaps there are coders out there that are happy to help code it up but just think that it's such an ingrained part of KDE they'd never have the clout to come in and shake everything up. If we can build up enough momentum here then perhaps we can get it done.

In fact I've just found a recent thread which pretty much sums up the situation for new users (and this guy even knows how to use a terminal!), entitled 'Im completely lost':

viewtopic.php?f=17&t=119102

Changed the title as the previous one wasn't really relevant.

I had some discussion with this on #kde-www a few months ago. The main problem is technical: if you want a replacement, alternatives need to be properly evaluated, and that takes time and energy.

It's also not something one single person can do: the KDE sysadmins worked together when they were evaluating a specific Git frontend for the repositories.

Hi einar. I understand completely what you're saying. At the end of day this is a major bottleneck from a visual design point of view which is why I have brought it up. Even the best looking themes are no good if the average user struggles to install them, or set them up properly. If visual design is to be taken seriously at KDE then this is a problem that needs to be solved, along with streamlining / combining all the different visual settings into a coherent blob in the system settings module, which is another area of confusion for the average user.

I can only hope this thread will maybe spark something off with the right devs who can start putting some wheels in motion. Because I know how quickly the months tick by, and before we know it we will be a couple of years down the line and still in the same boat.

The main problem as I see is that there's bound to be a lot of prep-work "under the hood" and this puts off potential contributors. Also so far we ("we" as in the wider community) have problems in attracting web people. Most that joined in the end left due to burn out or simply vanished after a while.

einar wrote:The main problem as I see is that there's bound to be a lot of prep-work "under the hood" and this puts off potential contributors. Also so far we ("we" as in the wider community) have problems in attracting web people. Most that joined in the end left due to burn out or simply vanished after a while.

Why couldn't the tasks be divided up into smaller pieces? That way perhaps parts could be written as GSOC projects for KDE.

davidwright wrote:...In the long run something has to be done, I think we all agree on that, so if we can at least gather some ideas here it gives us a reference point for the future. Perhaps there are coders out there that are happy to help code it up but just think that it's such an ingrained part of KDE they'd never have the clout to come in and shake everything up. If we can build up enough momentum here then perhaps we can get it done.

In fact I've just found a recent thread which pretty much sums up the situation for new users (and this guy even knows how to use a terminal!), entitled 'Im completely lost':

viewtopic.php?f=17&t=119102

That user experience is abhorred, and probably more typical than we would think. Also, too often KGHNS simply tells a user that the actual content they are trying to install is hosted elsewhere (DeviantArt, Dropbox, etc.) and off the new user goes down a rabbit hole they will likely never get out of. Something has to be done. The question, of course, is how.

But if it was done initially, it can be done again. And better.

Letting multiple GSoC students create a new subsystem collaboratively doesn't sound like something that would work well unless supervised extremely well.

If anyone wanted to work on this, the first bit would be probably just selecting a technical (pre existing) solution. That would be already a great help, given that OCS is in-house cooked and so any possible replacement would need evaluation.