Re: gEDA: The "scriptEDA manifesto"

["Piotr Miarecki" <sp9rve-at-NOSPAMpoczta_dot_onet_dot_pl>]

-----Wiadomość oryginalna-----
Od: Ales Hvezda <ahvezda@seul.org>
Do: geda-dev@geda.seul.org <geda-dev@geda.seul.org>
Data: 16 stycznia 2002 14:12
Temat: gEDA: The "scriptEDA manifesto"


    My private opinion:

    scriptEDA is a great idea. But it touches many aspects of integration of
different tools.

1. What is the integration ? It is possibility of using different tools in
the common work. It is necessary that one of them read without problems
results of work of other ones. For example: waveform viewer should read
results of work of SPICE, GNU Cap, Icarus Verilog simulation. They all use
(only) files to interchange data.

2. There are however some cases where scriptEDA (or other communication
channel) will be absolutely necessary: backward annotation, optimization,
advanced design tools. For example: if you make some changes in PCB layout
such as pin or gate swaping - it should be reflected in schematic diagram.
PCB tool has to send changes (or save them to file, then send command to
read it) to schema tool.

3. There are many ways to implement the communication channel. Using
scripting language is a comfortable solution and has many good points
mentioned in Timothy's mail. First of all a simplicity. It is however a
question numer 1. There are good known methods of implementing access to
data in other application (even remote): for example Corba. Why not to think
about ready to use, good specified and standard libraries ?

4. All gEDA tools that have scripting language use Guile. Why to reimplement
a number of tools to support Perl or Python ? What's more a mentioned Magic
uses Guile as well.

5. It is very dangerous to allow others to modify internal data in your
application in a direct manner. I think that scripts should be used only for
small and good defined class of operations. For example: no direct access to
structs in gschem for PCB tool, only a method for updating schema ! If
something is wrong, diagram is not updated (not consistent with PCB layout)
but also not destroyed.

6. I cannot believe that all developers will respect scriptEDA
specification. It is not necessary to guaratee that any tool would be gEDA
compatible (or how to call it ?). It is enough to define and implement
scriptEDA (or something else ) for gEDA tools and friends.

    Other question number 2. Is it right time to think of implementing
advanced features when not all base functionality is ready ?


Piotr Miarecki




-- 

Określ Swoje potrzeby - my znajdziemy ofertę za Ciebie!
[ http://oferty.onet.pl ]