Re: [Usability]File renaming/extensions



Hello again,

I've been thinking over this for some time and I've reached two
conclusions:

   1) Most of you who think that hiding extensions doesn't resolve
anything are right; it only hides the problem and impedes reaching a
proper solution.

   2) In order to reach some proper solution we had to have a proper
metadata system in place before anything gets resolved. Either at the
system level or at the GNOME level in order to support all platforms
GNOME supports.

I recall that Nautilus has a capable metadata system in place but I'd
have to have a look at it before telling if it is enough for handling
tasks.

Also the system should be powerful enough to make possible at least a
system as powerful as the old Mac OS with its creator/type tags. For the
benefit of those who don't have worked with it I'll try to describe it
at least in the surface:

   * Every file in the filesystem gets two tags that identify:
      1) The application that created the file, and
      2) The type of data stored in the file.

   * When you save a file with an app, the app tags the file with its
     ID and the data type.

   * When you open a file using Finder the file is opened with the
     application that saved it last, if available.

This system can, of course, be improved. But is should be at least a
good starting point to try to have its features in the first place.

Also extensions weren't needed because there was a database in place
that equaled file types with extensions that could be used in order to
send data to other systems (mainly DOS/Windows PCs).

You have also to take into account that most network protocols used
nowadays have bested the extension mechanism by sending the data type as
a separate entity to the data or the data file name in itself. Take
HTTP, SMTP and WebDAV as examples of it. When a Windows box with
Internet Explorer notices that the extension and the data type sent by
HTTP doesn't coincide the file gets saved by default with the extension
substituted with the one that is standard for the Content-Type header
received.

My final conclusion is that a proper system can be architected and
implemented without fear to compatibility. I'll try to do this and get
some documents going but if anyone wants to best me or collaborate with
me, feel free to do it, your efforts will be welcomed.

                               *   *   *

Also, in relation with Daniel's problem of having all the name selected
when renaming a file. My opinion is that the expected behaviour when you
select an object in a GUI is to have it all selected in the first place.
So the standard behaviour should be fine.

Now, what about implementing smart double click selection in file names?
The interaction should be something like this:

   User> Selects the file and either presses F2 or selects Rename from a
         menu.

   GUI> Highlights all the name.

   User> Doubleclicks or presses Shft-Ctl-Left to select all the 
         characters before the dot.




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]