On Plasmate's recent project list

[Shantanu Tushar Jha]

Hi,
    So on a whole I think we will not be forcing project and folder names to
be identical. So, how are we going to implement this? I tried creating a new
project "Hello" and changed it to "Hello1" in the metadata. Now, when I
again tried to create a project "Hello", it overwrites the older "Hello"
(renamed now to Hello1 in metadata, but the directory is still Hello). So, I
think when the new project with the same name is created, we should create
another directory (say,Hello-2 which is not visible to the user). Correct me
if I got it wrong :)

Btw, nice to see Plasmate gaining shape, nice work guys !! Though I couldn't
help much not knowing any scripting languages correctly. Still, it feels
nice :)

Cheers,

On Thu, Jan 28, 2010 at 2:01 AM, Diego Casella ([Po]lentino) <
polentino911 at gmail.com> wrote:

> ---------- Messaggio inoltrato ----------
>> From: Yuen Hoe Lim <yuenhoe86 at gmail.com>
>> To: plasma-devel at kde.org
>> Date: Wed, 27 Jan 2010 12:45:58 +0800
>> Subject: Re: On Plasmate's recent project list
>>
>> Correct, the project <---> folder naming convention is suggested, not
>>> required ( even though I wouldn't break that :P ).
>>>
>>
>> Hmm, okay so this brings up the greater question of whether we want to
>> force project and folder names to be identical (They don't have to be
>> technically, but we can force it programmatically). I'm personally not keen,
>> (I have in fact already broken that rule in the implementation) and here
>> again are my list of reasons :P
>>
>>    - Forcing the names to be the same has the benefit of neatness, but I
>>    don't think this is always desireable since we are allowing users to import
>>    existing plasmoids as well as download them from GHNS (eventually), and the
>>    user really has no control over the names of these external projects. It'd
>>    be pretty troublesome, at least if it was me, if I was blocked from
>>    importing just because I have a project locally with the same name (note
>>    that if we force project and folder names to be identical, name conflicts
>>    will occur even if I have a plasmoid and a runner, for example, with the
>>    same name, and those are clearly different).
>>    - I think importing existing plasmoids and stuff will be a fairly
>>    common use-case, and the project name == folder name rule is not widely
>>    enforced in existing plasmoids (my own plasmoid doesn't keep this rule..)
>>    - For the above reasons I'm not convinced that there is a significant
>>    advantage of forcing project and folder names to be identical, and yet
>>    forcing them to be identical will make a lot of other sticky
>>    conflict-resolution dialogs necessary, not just in this 'import-all'
>>    feature. Examples: the regular project import, import from GHNS or even the
>>    desktop if we implement that, and when the user changes the project name in
>>    the metadata editor.
>>
>> In short, I think forcing the names to be identical will create a lot of
>> extra work without really adding any significant benefit. It still can be
>> done though if you guys really think it's better. What do you guys think? :)
>>
>> So,if we bump into a conflict situation, we rename one of the two folder,
>>> good. Now, my question is: how the user will be able to distinguish the
>>> "current" and "backup" version of his/her project ?
>>> I mean, in the project list we can't show the directories name because
>>> they must be hidden, so an appropriate way is to pick up the project name
>>> from the "Name" field metadata.desktop file, and surprisingly this will be
>>> 99% times the same, since previously there was a conflict, so most likely
>>> the user will fill a bug report because he/she can't distingiush between the
>>> two projects, and he/she is forced to look to the sources in order to find
>>> the correct one.
>>> So, what about showing the "Remove,overwrite,ignore" buttons, or adding
>>> more informations in the project list (for example adding the date of last
>>> modification could be enough to distinguish between and old backup and the
>>> current project, at least when there are few projects). Any other ideas ?
>>>
>>
>> I maintain that the former only makes sense if we force project names to
>> be == folder names (in which case we'd need to add that kind of
>> options/dialogs all over the place). If we keep the current status though
>> (project names != folder names), then I agree that we need distinguishers in
>> the list. I was thinking adding the author name and version, because it
>> should be relatively uncommon for the same guy to create two projects with
>> the same name, so showing author should eliminate the larger class of
>> duplicate names that result from external imports.
>>
>
> +1 for me :)
>
> For people who actually want to maintain two projects of the same name and
>> both by me, version number I think is a sensible way for me to distinguish
>> between the two (so instead of doing the somewhat uncool thing of having to
>> name my projects coolplasmoid_1 and coolplasmoid_2, I could have
>> coolplasmoid v1 and coolplasmoid v2. Nicer IMO). Slipups that create same
>> plasmoid name and same author and same version can STILL occur, but that
>> would hopefully be rare, and again the fix is trivial - the user just needs
>> to load either one and check his code, then flip to the metadata editor and
>> key in an appropriate version number (or change the name if that's what he
>> prefers).
>>
>> Any other ideas? :)
>>
>> Yours is good and, since its not a vital component in our app, we could
> change its behaviour later, referring on users feedbacks :)
>
>
>> ----
>> Jason "moofang" Lim Yuen Hoe
>> http://yuenhoe.co.cc/
>>
>
> _______________________________________________
> Plasma-devel mailing list
> Plasma-devel at kde.org
> https://mail.kde.org/mailman/listinfo/plasma-devel
>
>


-- 
Shantanu Tushar    (UTC +0530)
http://www.shantanutushar.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.kde.org/pipermail/plasma-devel/attachments/20100204/63acbd1c/attachment.htm