Additional shapes ways of use and distribution [+digest try] [was: Re: Diashapes 0.2.2 released]



[thread renamed as it's not anymore the initial subject concern]

Le 28/08/2010 07:43, Steffen Macke a écrit :

Let's try to assess how many more users this will enable to install
additional shapes:

* It's the difference between the number Debian users that would be able
to copy the contents of a zip file to the right location and the ones
running Debian sid

My guess is that this number is close to zero, because Debian users are
clever.


[please forgive my poorly written English]

* sid is not the better way to distribute "temporary packages": for that
kind there is a special debian repository I currently don't remain
the name (I used it in the past only for clamav virus database,
which is now an integrated cron job in debian).

* a cron job would be far ever better.

* Compare this to the number of users running Dia on Windows able to
deploy the zip file contents vs. the number of users being able to
operate Diashapes (should match the number of Dia users on Windows).

* at is not cron ;)

In this case my guess is that the number is a high percentage of the
number of Dia users on Windows.

Alternatives: Just write this little tool that plugs Microsoft update
into Debian sid for this purpose :-)

Hope you don't talk MS update(r) but dia shapes updater itself.


To be en par with Diashapes in terms of granularity, of course, you
would have to package every single sheet in it's own package - it that
what you would like to do? If that's the case, I should split up the
dia-additional-shapes package into individual sheets.

Well, this is a complex case :
 * Some debian installations are necessary never plugged to the Ternet,
    even never plugged to any network (I know that: I'm in the forces).
 * for Dia itself, Dia installation, even on M$Win, can be restricted
   to   unplugged computers ("legal" security concerns which needs to be
   addressed a complete security review/tests by internal entitities
   security teams)
 * as said in the thread, Sharing/loading sheets and shapes
   worldwide would be great, but could dramatically increase the number
   of shapes/sheets, with quality concerns
   (I spent ten of hours simply to re-design a "modern shape standing
    for 'desktop computer'" which re-sizes nicely, with a graphical
    svg editor (now named Inkscape) plus "vi" to correct svg
    connections then add shape specific stuff
-- http://harding.fr/tom/graphics/dia-shape/downloads/desktopprinter.shape --
   ).
 * Initially, Dia Diagrams were mainly used by sys/net admins, and
   a fe


######
Intro
######
I had been involved in Dia devel for a few months, focusing on
documentation building an it's I10n processes. I also sent a
little few "minimal" shapes which are now integrated, at least
in debian 0.97.1 Dia version.

I have no capability in C/C++ except shorts hacks (I prefer to let
others do the underlying process then use high-level languages, such
as Python), for example : fix slrn for newsgroups moderation purpose,
change a library call in Python acl library (then send patch and idea
to upstream), propose a try on a "tree shape" to dia upstream.

I'm proud my little contributions effects to Dia, which takes me months,
had been integrated, then carefully corrected by others, improves Dia
itself at least for my own usage, and usage of Dia by a few un-numerable
persons : I didn't deserve a tool I use sometimes myself, by matter of
need, and both a little bit proud and a little bit annoyed each time
someones says me the thanks for the outdated French doc translation
which continue to figure on the Air French Force free software
repository. I didn't ask to drop out the thanks because they mean for
me : you can improve, and you can share.

And there is also a sheet produced by French forces about physical
security which seems Dia distribution discontinuated. Maybe because
of it's specificity (note I never endorse or participate on that:
I show here a sample of somewhat complex is the process of a sheet
sharing).

Dia is not a general purpose drawing program, it is a tool addressing
diagrams making purpose. A diagram is a semi-graphical description
of something, which mean a level of an basic and general to a very
specific and precise abstraction, by a combination of what we call
in french "logos". Ancient Greek "logos" means "language". A "language"
involves "Babel Tower problem", such as context. Sheets addresses
contexts, multiply the sheets is an addressing of diagram contexts
for an increasing number of purposes.

I read "comics" since my young. "comics" are not limited to "comics"
but can have multiple purposes, from novel to essay, historical, technical documentation, comic strips, tracts, and any purpose
someone can draw or write a "scenario".

Comics are a "media" (at least in French end Belgian). Like diagrams.

Dia itself is not only a tool: /Dia is a media creator tool/.

This imply that Dia maintainers cannot guess what purpose any user
can address by using Dia.

In fact, to build then use a "good" Dia shape/sheet is a complex
process to reach "an efficient gain, by addressing /a mean/".

This also imply Ternet-Wide and possibly sites-wide sheets/shapes
distribution needs "moderation", without it Dia usage should
certainly become a M$visio-like nightmare.

As I said, I have no more possibility to help in Dia software
improvement, excepting rare and little contributions and
/thinking of what could be done/. In fact, my contributions
costs against Dia development and maintaining, AND anyone
contributing to any Free Software should be aware of that fact.

So, let users be aware the heavy pound their little contribution
can have against their tool, and let sysadmins and developpers
to moderate not well thunk contributions is the way to a good
practical /graphical languages/ expansion.


#########
\end intro
#########


############
proposal
############

So, let me a try on how to not geopardise the Dia use in it's
way of "simplicity":

 * to let Dia "basically" usable, initial amount of sheets/shapes
   /must/ be low (contrary to MSVisio nightmare). The "basis"

 * If we want to keep sheets/shapes usability, the "frequency/last"
   current menu behavior should be modified to comply with a large
   amount of sheets. So, a "purposes" entry could be added :

      +---------------------------
      |Basic sheets [edit][reset]:
       (thinks edit will address "basic" usage for
          any person)
      | misc
      | network
      | xxxx
      | xxxx
      +---------------------------
      |Last:
      | xxxxxxxxxxxxx [del]
      | xxxxxxxxxxxxx [del]
      | xxxxxxxxxxxxx [del]
      | [+...] (thinks should be limited to 5 or 6 last entries)
      +----------------------------
      | Purposes:
      | philosophy [del]-> are you sure want you loose your own setup?
      | [add]
      +----------------------------


       Where [add]or [edit] Purposes will conduce to a tree-like menu,
        modeled after debian-like "tags" into sheets description:
          debian "tags" allows multiple tree entries for the same
          program, so "tags" will allows multiple menu entries for
          shapes and sheets.

 * the free software spirit of Dia can be extended to "usage" by
   "shape sharing". this "spirit" means:
     * create a shape editor mode with non-svg shape specific stuff
        (connexion points, editable text node, and so on)
     * let users licences aware ((L)GPLvX,BSD2,Artistic), give them
       a convenient choice with an explanation of each (limited choice
       of) proposed licenses.
     * force users to "tag" by proposing the complete set of tags, plus
        an "add tag" entry.
     * "moderate" the users sheets/shapes entries in the
        "shapes repository".
           => "purposes" "tags"
           => sheet and shape names and I10n
           => shapes quality
     * create a "shape repositories URIs setup" menu entry
     * for non-shared shapes, integrate an option to share shape
        in dia file itself (which means to implements svg "viewport"
        to make it usable, and dia file retro-compatibility lost.
        This lost is not a new problem, because of the prominent need
        to have the not yet integrated shape is the main problem
        to display, edit or export the dia file).
     * for a site-wide dia installation, having non-standards shapes
        recorded by default in /var/lib/dia/shapes/ or "all users"
        by default instead of user's home would be the good trill.
        Note I'm not a M$Win administrator (never will I ;), so
        I'm not sure about "all users" profile.
          * This should not replace the "record non standard shapes
            together with diagram into dia file": mail is not nfs
            transparent.
      * build nightly the whole tar/zip archive the whole shapes/sheets
         repository for unplugged/isolated networks purposes.
      * offer a distributable web search engine with tar/zip building
         of selected sheets and output.



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