Mail Archives: geda-user/2016/02/28/17:51:01
--Apple-Mail=_0EBF1587-01F7-4DBD-B323-661DD2D6A30E
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
charset=windows-1252
On Feb 28, 2016, at 3:22 PM, Britton Kerin (britton DOT kerin AT gmail DOT com) =
[via geda-user AT delorie DOT com] <geda-user AT delorie DOT com> wrote:
> On Sun, Feb 28, 2016 at 11:50 AM, John Doty <jpd AT noqsi DOT com> wrote:
>>=20
>> On Feb 28, 2016, at 1:10 PM, Britton Kerin (britton DOT kerin AT gmail DOT com) =
[via geda-user AT delorie DOT com] <geda-user AT delorie DOT com> wrote:
>>=20
>>> On Sun, Feb 28, 2016 at 6:11 AM, John Doty <jpd AT noqsi DOT com> wrote:
>>>>=20
>>>> On Feb 28, 2016, at 7:07 AM, Kai-Martin Knaak <kmk AT familieknaak DOT de> =
wrote:
>>>>=20
>>>>> John Doty wrote:
>>>>>=20
>>>>>>> as well since it will help gschem stay viable long-term,
>>>>>>=20
>>>>>> Gschem has no such problem.
>>>>>=20
>>>>> CERN chose kicad rather than gschem for their open hardware
>>>>> initiative. Same with a number of high profile open hardware =
projects.
>>>>> See http://kicad-pcb.org/made-with-kicad/ for some examples.
>>>>=20
>>>> And more people choose Word than LaTeX. That doesn=92t mean that =
LaTeX isn=92t a better choice for those of us who use it, or that it =
should be more like Word (at your suggestion, I tried Lyx. Ugh. =
Inflexible and difficult. Combines the disadvantages of both =
approaches).
>>>>=20
>>>> Eleven days ago, JAXA launched Hitomi (ASTRO-H) with ASICs designed =
with geda-gaf and ngspice on board. Quietest CCD measurement chains ever =
put into space.
>>>>=20
>>>>>=20
>>>>> Please don't delude yourself int the perception everything is =
working
>>>>> alright, when it clearly isn=92t.
>>>>=20
>>>> Clearly not. But there=92s a profound difference between what pcb =
users want and what geda-gaf provides. Geda-gaf's application space is =
far broader, but pcb users want something more targeted. Of course, =
geda-gaf is pretty simple and flexible, and we who script it can aim it =
at whatever target we wish. But there's a big faction here who don=92t =
want to do that: they conceive of something as hard-wired, complicated, =
and inflexible as pcb (but perhaps easier to use for *their* =
application).
>>>=20
>>> pcb has an action system. It has batch mode. It has a plugin =
system. It has
>>> a text file format. It doesn't aspire to be "hard-wired" any more =
than gschem.
>>> A number of people are working on making it more scriptable, not =
less so.
>>=20
>> But it doesn=92t have flexible abstractions underneath these things. =
Pcb simply can=92t do anything that fails to conform to a limited set of =
intentions reflecting use cases. It=92s not a =93stack of materials=94 =
editor: it can only handle a weirdly restricted subset.
>=20
> If you mean it can only make pcbs you're correct, but since that's all =
it's
> supposed to do that's hardly surprising. The abstractions it uses are
> quite suitable for the purpose.
No they are not. Suitable abstractions would allow the user to draw a =
buried via, for example.
> If it was a "stack of materials" editor
> it wouldn't work to make pcbs, since stacking up a bunch of materials =
won't
> result in a pcb.
But PCBs commonly contain structures that pcb cannot represent, so pcb=92s=
model is deficient.
>=20
> I think what frustrates me most about your attitude towards pcb is =
that you're
> determined to regard its problems (and I freely admit that there are =
some)
> as fundamental design defects. You're mostly wrong in this, have no =
basis
> for the claim, and should stop making it.
>=20
>> Gschem is an unspecialized topology editor built on highly flexible =
abstractions. Add a suitable symbol library and it can represent the =
topology of a printed circuit, an ASIC, a Makefile, or plumbing. There =
are a few blemishes, e.g. slotting, where it hard-wires use cases, =
pcb-style. Perhaps we can move those to optional scripts in the future.
>>=20
>>>> It=92s also true that much of geda-gaf=92s intelligence is actually =
in your symbol library, so assembling a good symbol library for a =
particular purpose is a large part of the challenge. But as you know, =
this is also fraught with difficulty, because even the pcb application =
space isn=92t a coherent, monolithic thing.
>>>>=20
>>>>>=20
>>>>>=20
>>>>>> It has been stable and productive for
>>>>>> well over a decade.
>>>>>=20
>>>>> Just like pcb. Only that in this case it is decades.
>>>>=20
>>>> Except that if you watch this list, the most common topic is =
problems with pcb. A significant subset is problems with pcb that people =
expect geda-gaf to work
>>>=20
>>> That's because what pcb is trying to do is harder.
>>=20
>> Is it that the problem is intrinsically harder, or that the approach =
of implementing a function for every use case ad hoc makes the problem =
harder?
>=20
> It's intrinsically harder. I'm going to assume you accept that.
I do not. I accept that you can make it harder by having an inadequate =
foundation.
> If you
> knew how it worked at all you wouldn't claim that it implements a =
function
> for every use case. It absolutely does not. If you actually want an =
idea
> how it works take a look in src/gpcb_menu.res.in, and you'll see that =
all
> the operations defined in the menus are implemented in terms of =
sequences
> of fundamental actions.
You have a different notion of fundamental than I have. What you call =
fundamental, I call ad hoc. And the proof that I=92m right is that your =
so-called =93fundamental=94 operations do not actually capture the =
construction of a PCB in general.
>=20
>>>>> Do you remember the poll I did a while ago? The results were non-
>>>>> ambiguous.
>>>>=20
>>>> No, they were on a pcb-dominated forum. Not an unbiased sample.
>>>=20
>>> The forum in question is also the largest gschem forum
>>=20
>> Indeed. That=92s a problem. We really need to separate the projects. =
What=92s here of interest to a gschem user who doesn=92t use pcb? People =
like Paul Tan don=92t seem to post anymore.
>=20
> You're changing the subject from the user composition question and =
relevance
> of Kai's poll. I'll take it you concede the point.
Of course not. You=92re not being scientific. It is certain that the =
population here is not representative of the whole community, and there =
are strong reasons to believe that geda users who don=92t use pcb are =
underrepresented. You=92ve worked hard to make us unwelcome.
John Doty Noqsi Aerospace, Ltd.
http://www.noqsi.com/
jpd AT noqsi DOT com
--Apple-Mail=_0EBF1587-01F7-4DBD-B323-661DD2D6A30E
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename=signature.asc
Content-Type: application/pgp-signature;
name=signature.asc
Content-Description: Message signed with OpenPGP using GPGMail
-----BEGIN PGP SIGNATURE-----
Comment: GPGTools - https://gpgtools.org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=Egap
-----END PGP SIGNATURE-----
--Apple-Mail=_0EBF1587-01F7-4DBD-B323-661DD2D6A30E--
- Raw text -