Hi,Thanks for the response, all constructive criticism is welcome ^^. This is also due to the fact that my user interface proposition was a UI proposition only - and nothing more. This means that this issue is still open and nothing is known for sure.
I would like to comment on Waldemar's response first. Waldemar Kornewald wrote:
If we get a much better design then I think radical changes are OK. Let's also consider creating an installer UI that is simpler than the BeOS one. If we don't like it we can still have your suggestion (needs only minor fixes for compliance or wording).
That's of course true, but under the assumption that the resulting pros are really significant. In my opinion, this is a very important issue to discuss, since depending on the decision our field of action is changing.
Maybe instead of "<Package name>" the window title should be "Install: <Package Name>" to make it easier to identify?
I agree. This way seems better, since then an unexperienced user knows exactly that he will be installing, not running the given application/package.
I'd change the label "Groups" to "Installation Type" or something similar.
I agree as well. I thought of this too, but more in a way like 'Installation Profile'. But 'Installation Type' seems better, since it sounds more obvious for newbies.
I think the tabs view should start at the same height as the list view (instead of its label) to make it more obvious that both views are connected. The list view's label could be replaced with a BBox to make this even more obvious.
I personally don't like this idea too much. That's because I tend to feel that tab names are equivalent to labels of subsections, and this IMO would make the application look chaotic. As for the BBox, I think this shouldn't be necessary. The list view has its own border anyway, so another one from the BBox would make it look busy, again.
I like that you replaced the alert after pressing Begin with a simple tab, but OTOH, do we really need that info, at all? It's still available in the log files. Isn't it sufficient if you just select the target location and that's it?
Even though most target users won't be interested in such, I think we shouldn't hide important information like this from everyone just because of this. That's why I proposed the tab view with the 'Files' being the second, hidden by default, tab - so that it wouldn't confuse users too much. This is, of course, only one possibility. Instead, we could, for instance, include a button 'Show files' (or similar) that would pop up a new window with the tree.
As far as I remember, many apps only have one single profile and there are rarely more than three profiles. IMHO, a list view is only appropriate if: * you expect to have many entries * switching between entries must be very efficient * you must have an overview of all elements I think that of those three only the last one might apply, but a pop-up field could be better suited for this task and result in a more compact UI.
This makes sense. If, of course, we decide that similarity to the BeOS version is not of high importance.
Is there any need for splitting the installation volume and the installation folder ("Install on ... in ...")? Why not have one pop-up with the full path? The pop-up could automatically offer pre-defined paths for every volume and a "Custom..." item for choosing your own path (when paths can't be chosen "Custom..." would be grayed-out).
That could be a good idea. The only thing against it is that most normal, desktop users don't care where exactly is the application installed. The only thing interesting is the volume (or, we should say, disk) to be used. Although I personally see no big usability differences in these two approaches.
An alternative installer UI could look like this:
Personally I like the interface you proposed - with a few remarks of course. For example, I think adding a 'Show files' button would be very convenient in this approach.
This being said, thanks for your opinion. I think all propositions should be discussed before taking any action in this field. The most important thing I would like to resolve before anything else is the issue of similarity. Should we or should we not make it as much similar to SoftwareWallet?
Since I'll probably be absent for the whole day today, I will continue my thoughts tomorrow. Till then, keep the discussion going! ;)
-- Best regards, Łukasz 'Sil2100' Zemczak