Mail Archives: geda-user/2016/07/24/05:01:58
Britton Kerin (britton DOT kerin AT gmail DOT com) [via geda-user AT delorie DOT com] wrote:
> On Sat, Jul 23, 2016 at 8:40 PM, Vladimir Zhbanov (vzhbanov AT gmail DOT com)
> [via geda-user AT delorie DOT com]<geda-user AT delorie DOT com> wrote:
>
>> On Sat, Jul 23, 2016 at 05:35:07PM +0200, Bert Timmerman (bert DOT timmerman AT xs4all DOT nl) [via geda-user AT delorie DOT com] wrote:
>> ...
>>
>>> John Doty is right in that gschem will not lock you into pcb.
>>>
>>> gschem and gnetlist do give *you* the choice where to go for *your* next
>>> step in the workflow *you* want to follow.
>>>
>>> pcb is just one possible further step, and osmond, verilog, vhdl (ghdl),
>>> gnucap or ngspice are other possible steps (see "man gnetlist").
>>>
>> While I agree with you here, there is another side of the
>> coin. There are netlister's backends nobody uses and nobody
>> reports their state. A few months ago one user has reported vams
>> is broken. Looking at it I've discovered it is broken about a
>> decade ago. No special tests were made at the time of its creation
>> so just nobody knows...
>>
>> I am saying about this because my intention, which I am going to
>> implement, is stripping our repository by removing various
>> obsolete/unused/non-working thigs. Of course, I'll ask our users
>> here about any of them, if/when I do this.
>>
> Good idea. Things can always be brought back, but as it stands
> users have to guess which are the functional bits. Same goes for
> parts of some of the packages, e.g. gscheme hides it's useful
> general schematic components amidst a random smattering of
> obsolete vendor-specific crud.
>
> Ideally the web site/repo/whatever should reflect as accurately
> as possible the state of its components in terms of user base
> and testedness. This is especially important for minority
> software like gEDA where fear of being alone can deter users.
>
> Britton
>
>
Hi Britton and list members,
Just from the *pcb* point of view, are there any ideas how this could be
achieved for pcb ?
As a (generated?) list or table of features [HID's, exporters, utils,
scripts, etc.], landpatterns and their status ["under construction",
"testing", "released for production (vetted)"], and the number of users
using these features/landpatterns, and the number of active bug reports
on a feature/landpattern ?.
Maybe we would need some sort of poll (or some other mechanism) to
figure out what gets used the most.
And then need to keep it up to date !
I like to hear your ideas on this one.
Kind regards,
Bert Timmerman.
- Raw text -