[Kexi-devel] Use Case; was: KDE (vs GNOME)

[Jarosław Staniek]

Gianni Rossi said the following, On 2005-11-14 15:06:

> On 11/14/05, Jarosław Staniek <js at iidea.pl> wrote:
> 
>>Gianni Rossi said the following, On 2005-11-14 14:00:
>>
>>
>>>On 11/13/05, Jarosław Staniek <js at iidea.pl> wrote:
>>>
>>>
>>>>Cornelius Schumacher said the following, On 2005-11-14 00:52:
>>>>
>>>>
>>>>
>>>>>On Monday 14 November 2005 00:26, Jaroslaw Staniek wrote:
>>>>>
>>>>>
>>>>>
>>>>>>-Since certain database backends may require latin1 identifiers, these
>>>>
>>>>are
>>>>
>>>>
>>>>>>created implicity for level 1 (user do not know about this).
>>>>>
>>>>>
>>>>>Why would any user want to explitcitly give identifiers when they can
>>
>>be
>>
>>>>>created implicitly?
>>>>
>>>>See my answer to Robert Knight for more info.
>>>>
>>>>
>>>>
>>>>>This doesn't sound like a good reason for user levels,
>>>>>but as a good reason to remove some complexity from the user interface.
>>>>
>>>>To fully understand this need, you need to look how database
>>>>tables/queries/
>>>>etc. are named in a real-world database (or data warehouse)
>>
>>applications:
>>
>>>>e.g. you can encounter very descriptive names like "Annual report for
>>>>medicines" table name. One user may want to use this long the name in
>>
>>SQL
>>
>>>>statements and scripts, other do not want this and want to just put his
>>>>own
>>>>abbreviation. He/she is the developer in this case.
>>>>In the same time simple users should not be even aware of above issue if
>>>>they
>>>>do not want. They most probable wouldn't want that unless they are using
>>>>SQL
>>>>or scripting.
>>>>
>>>>We _cannot_ assume everyone is simple user nor advanced user. I
>>
>>mentioned
>>
>>>>about data warehouses. I don't know about your impressions, but in this
>>>>case I
>>>>encountered cases when I already have predefined latin1 names in one
>>>>linked
>>>>database (ofter only 8 or 12-character long) and just in my master
>>>>database I
>>>>wanted to add descriptive captions that I can then use in GUI and
>>
>>project
>>
>>>>documentation.
>>>>
>>>>Another place power users may want to see real names and not captions by
>>>>default is project navigator ( on the left hand:
>>>>http://kexi-project.org/pics/0.9/data_form_navigator.png ) or
>>
>>table/query
>>
>>>>column names.
>>>
>>>
>>>
>>>Just a question. How would an advanced user "communicate" with a normal
>>>user? For example, one user creates a table using captions, other uses
>>>names, or vice-versa? I'd think that to reduce confusion, BOTH names
>>
>>should
>>
>>>be seen all the time, regardless of user status; though, the capton/name
>>>should be put to a side in smaller print or something. I'm thinking
>>
>>about
>>
>>>the case where there are many tables/queries with simmilar name and an
>>>advanced user is talking to a normal user.
>>
>>Good question. Problems with communicating between "real database
>>developers"
>>and newbies is not so depenent on our idea of user levels:
>>
>>- Look that if caption is not available, Kexi falls back to latin1 name
>>everywhere in the GUI (in particular, for table and column names)
>>
>>- always displaying latin1 name to newbies does not mean they will choose
>>reasonable short name that advanced user may prefer. Thus it can be better
>>to
>>stay with automatically generated names so advanced user at least will see
>>what to expect from a project created by a newbie.
>>
>>In short, by hidding the name I have chosen to be as similar to other
>>popular
>>tools (MS, Filemaker) as possible. The target audience comes from
>>spreadsheet
>>world, so it's a priority to hide technical storage-related stuff by
>>default.
>>
>>One thing we could added to the dialog
>>http://iidea.pl/~js/kexi/shots/0.9/object_save_as.png
>>for the 'newbie level' is a "specify name" checkbox or "advanced >>"
>>button.
>>What about this?
>>
> 
> 
> That seems ok. It would be nice however, if kexi sugested a default
> "meaningfull" name based on the Caption; ex:
> 
> Capiton: Office Address
> Name: OFFICE_ADRESS
> 
> or
> 
> Caption: Situação Temporária
> Name: SITUACAO_TEMPORARIA
> 
> Notice the removal of "çãá" in favour of "c" and "a"

This is already implemented and works well. Of course only for languages based 
on roman alphabet (or allowing a conversion to such).

> But, then again, I guess that would be language dependant... how would this
> work?

I have got special conversion table, see string2Identifier_conv table in 
kexi/kexiutils/identifier.cpp. AFAIK thee's no such table available yet (not 
mentioning in KDElibs) so I am maintaining own impelemntation.

BTW, I recently received good conversion for German. There's also polish 
conversion, and that's all.

Could you (or anyone) provide similar conversion table for your language 
(group of languages)? Many of them can be even guessed without a need to know 
the language.

-- 
regards / pozdrawiam,
  Jaroslaw Staniek / OpenOffice Polska
  Kexi Developer:
      http://www.kexi-project.org | http://koffice.org/kexi
  Kexi support:
      http://www.kexi-project.org/support.html
  KDE3, KDE4 libraries for developing MS Windows applications:
      http://wiki.kde.org/tiki-index.php?page=KDElibs+for+win32