List of propositions for gEDA-versions to come

Bernhard Kr?mer wrote: : Hello,

: After two days of desperate tries to make one simple circuit with gEDA, I am : nearly giving up. In contrast to Protel, where the functions are hidden, : gEDA is much better organised, but also much harder to use, especially for : a novice.

: At this point, I propose the following changes to be made in future versions : of gEDA. I know that this is perhaps not the best place to post such a : list, but I hope you still appreciate it :

Constructive criticism and ideas are always of interest!

: - The menus should be steerable by keyboard. They should disappear when : clicking beside of it.

If you mean that the menus pop open when you type the corresponding keyboard shortcut, this is an interesting idea.

: - The footprints should be assigned for many components at once, by : selecting them and choosing one footprint which will be then assigned for : all.

Use "gattrib" to do bulk editing of footprints and other attributes. It presents your design as a spreadsheet components and their attributes. Here's a screenshot:

Don't download gattrib from that site -- use the gattrib version off the CD.

The issue with footprints comes up often, especially with newbies.

The gEDA philosophy is minimalist w.r.t. footprints and other attributes. That is, the component symbols aren't pre-loaded with footprint attributes. It's up to the user to make the assignment of the footprint names. This is the case for several reasons:

• Each user may have his own names for the footprints.

• Many common parts have multiple footprints for the same schematic symbol. Examples include most passives (e.g. 1K resistor in through-hole version, SMT of various sizes). Instead of having one symbol for each 0402, 0603, 0805, 1206, etc, you just assign the footprint you want at design time.

• It's just evolved that way. Building and maintaing symbol and footprint libs is not glamorous work, so nobody volunteers to do that. INstead, folks using the tools just get used to rolling their own. That's what pros do anyway -- they roll their own footprints and symbol at least 80% of the time.

That having been said, one member of the gEDA community was going to produce a CD containing a professional quality symbol and footprint lib, and then sell the CD. He seems to have dropped the project recently. In any event, the question is: how much would you pay for such a product anyway?

: - Every component should have a footprint by default

See above. Defaults are good for newbies, but bad for experienced folks who have their own footprint libs and don't want the default values. Yes, defaults can be replaced, but you can get bitten if you forget to do the replacement. With no default, the DRC checker tells you to install a footprint.

: - When adding a footprint or whatever, standard should be "show value" and : not "show name&value"

INteresting suggestion. I agree. I can look at this.

: - There should be a footprint-browser in gschem. I could imagine a list of : possibilities that appears while choosing a footprint for a component. This : would spare you to know all footprint-types by heart and even worse, to : have to search for them.

This idea has been suggested before. It's a good one.

: - The preview option in the part-browser should be activated by default.

I can fix that.

: - After changing the schematics, new parts should appear in the pcb-file and : not in the .new.pcb-file.

Yeah, I'm not sure why it was architected to work like this. . . .

Stuart

Stuart,

I think you make a lot of good points regarding configuring gEDA for newbies vs. experienced users. I've found that -- with some effort -- software can often cater to both users just about equally. My comments (not all of them reflect this particular comment):

If you integrated that into gEDA itself, it'd probably make everyone happy (...and look a lot like ORCAD!). All the schematic capture/PCB layout programs seem to have their own 'paradigm' for performing bulk editing, but in general all the commercial programs can do it within the program itself. (In fact, one of Protel's selling points was that ALL the 'editors' ran without one window -- that might be going a little overboard, but I can certainly see the appeal).

$39.95? :-) Probably not worth the guy's time, but still, if he's planning on putting in the effort anyway, any extra income is nice.

So have a checkbox labeled something like "Substitute generic footprint for parts without them," (and perhaps have the DRC checker issue a warning rather than an error when a part doesn't have one).

Why not just get the preference to be remembered? Have a checkbox for "make these the default preferences" each time the dialog comes up.

---Joel Kolstad

I don't like to seem negative, but this naming business is a big failing of the Linux community. Instead of a descriptive name that goes some way to telling you what it's for, there's a tendency for names that are purely functional (after all, everything's got to be called something) or cleverish acronyms that can be disastrously misleading. I recently had a problem over the tip terminal program. Needed to use it, but could I find any documentation for it? man tip with my installation produces no results, and a web search for 'tip'... well, guess.

Paul Burke

Stuart Brorson a écrit :

I am very pleased to hear that! So you are a programmer who really cares about his baby ;)

Actually, it wasn't that what I meant. In fact, I am used to menus as everywhere in the Linux and Windows world: For example you click on "Size" and the menu "Size" appears. If you realize that instead of "Size" you wanted to go into the menu "Edit", you click on "Edit" and "Edit" appears immediatly. This doesn't work yet. Also, if you realize that you didn't want to go in the menu at all, you click anywhere else on the screen and the menu should disappear. What I meant with "steerable by keyboard": Yes, it might be interesting to open it by keyboard shortcuts. Furthermore, it would be very nice to steer in the menu with the arrow keys.

Honestly, I like the gEDA interface very much for its speed. I can use gEDA in the same way on my old Pentium I Notebook as on my Athlon Palomino. But on the other hand, the way the menus and so behave is outdated and very difficult for novices and for occasional use. I think this is a point where many users turn away from gEDA. This and the fact that the footprint system is still based on the old m4-system.In my layout, all the polarised components have been reversed and I think it is because m4 is no more maintained correctly. Such a thing /should never happen/ and if I could contribute a bit, being a non-programmer but interested in well working Open Source Programs, I would do it.

So my question: Could it be possible that all motivated gEDA-users conceive a roadmap for a new, modern interface and symbol/footprint system, and then program it while keeping all useful, already programmed internals? So instead of hacking around it, they would create a great, integrated Electronic Design Software. The base for that project exists already with gEDA. Compared to professional software like Protel, gEDA can do almost the same things and often even better. If the interface would be modern and intuitive and if the footprint/symbol system behind would be coherent, why should they still use other software?

Fine thing.

That is all true. What do you think about the following: First, you install gEDA. Every part has a default footprint. But you always can attribute a new default footprint.

I don't know if I would pay for the CD if all the software would be free and especially if I am an occasional or a new user. And then, if I was a professional every-day user, I agree that then I would have create my own symbols. No, I think that the geda-symbols-repository is a much more valuable thing.

Best,

Bernhard

tip is usually spelled: kermit on most linux machines ;-) I don't think I have seen tip since I used to work on sunos machines.

I am pretty sure that you can directly use the standard xterm window by piping something (cat maybe?) to and from the appropriate /dev/ttyS.

I've always used kermit.

-Chuck

Would it fall under the GPL (or LGPL)? Seems to me, you're right about the as-long-as-you're-doing-it-anyway thing.