Kicad - GUI is Horrific!

[Simon]

I haven't used kicad since 2011 and i dont remember editing a part causing thr netlist to break.

So im curious did you change the pin numbers on the part or did you simply move them?

You're coming in late and missing the point.  The result of my editing a symbol was the revelation
that schematic capture has no netlist.  The lack of a netlist in schematic capture is the point.
KiCAD only generates it later, as a postprocessing step from the completed schematic.

why are you so hung up on the netlist? file -> export -> netlist, you can even have it in a choice of formats. I am so glad that Kicad has gotten over the separation of schematic and PCB layout to the point that there is a whole step of netlist and component to footprint association to go through.

You may try to calm down a little. I have found the GUI one of the easier ones to use with the keyboard shortcuts making it a dream to use. For a wiring loom schematic it is ace, or was, they seem to be going through that phase that even professional software goes through where they have to ruin stuff first. Yes they are up their own assholes a bit, sadly that is open source, bitching to death about it on here won't solve it.

For example you used to be able to simply mouse over an item and hit "c" do duplicate it. Firstly they removed this, then you discover that reassigning this shortcut under the schematic does not make it work again, apparently this is because that is for the symbol editor, yes it is nuts. The solution is that this shortcut must be set under the general category.

Yes the person that dealt with my technical support complaint about this is a jumped up little fucking asshole as i was told that this is a feature rather than a bug basically. But it's free so what can I do about it? even when I used professional software (Circuit studio) the response of support was always that it was a feature not a bug.

I left Kicad on version 4 when it was just a mess and spent £2000 on pretty mush most of the cheap commercial packages only to find that they were all junk to me and I ended up back at KiCad even though it is not perfect it is a damn site better than any of the other entry level programs where they basically take your money then tell you to go fuck yourself!

This asshole attitude is something you will get in many open source programs, I run BOINC, I don't dare go on their forum anymore to ask for help or point out problems, the responses have ranged from someone claiming to be a developer being abusive to my posts simply being deleted despite being not at all controversial. Currently I have this problem where I need to set more than one time slot to suit my energy tariff, as there is only one slot available I often forget to change it and an hour or more of time computing may be lost, pointless asking about this so I have not bothered.

[ebastler]

As pointed out earlier in this exchange, propellerhead (under the user name "via") has already brought this up a few months ago on the Kicad forum. We can save ourselves time and aggravation, since that thread has fully explored the matter and ends with a very nice summary by moderator eelik:

I agree with Via that not handling connections as connections but as graphics is somewhat problematic and I would prefer having some way, whether it be only perceived or actually reflecting the architectural implementation, of preventing accidental de- and re-connecting implicitly something while modifying a design. But I also agree strongly with Jon that the tone in this discussion haven’t been constructive, whether Via feels that way or not. I honestly have felt some arrogance here.

I certainly have my disagreements with many design decisions in KiCad implementation or behavior and have certainly annoyed the developers more than once. Yet this isn’t a good way forward. I’ll close this thread for now. The facts and opinions have already been expressed anyway. The subject matter itself is open for discussion in the future, of course.

https://forum.kicad.info/t/symbol-editing-and-the-netlist/39271/110

[propellerhead]

why are you so hung up on the netlist? file -> export -> netlist...

Simon:

Yes, of course, once your schematic is done the netlist gets exported and moved through
packaging, etc.  As I said, that's not the point.  The point is that while you're entering the
schematic there is no netlist (that is, there isn't one being created and maintained by
schematic capture as you work), and doing it this way makes the schematic entry more
error-prone.  As others in the conversation have noted, there are examples of both methods
out there among the different packages, but in general, the higher-end, more mature, and
(incidentally) more expensive ones maintain and enforce the netlist in real time because
there's significant value in schematic capture understanding the difference between things
that are connected and things that aren't, and KiCAD has no such understanding.

Based on many years with an old version (V8) of Mentor Graphics, I'm suggesting that this
would be a worthwhile addition.  The problem is pushback from those who think it's purely a
matter of "opinion" rather than the consensus of people who really are (and I'm not including
myself among them) EDA experts.  In any case, the response is baffling, because they're arguing
for a tool that does less work for them and makes it easier for them to drive errors into their
designs, apparently because they think it's right because that's how it's always been done here.

If you care to go back a couple of pages I offer examples as to how this works.

[nctnico]

I haven't used kicad since 2011 and i dont remember editing a part causing thr netlist to break.

So im curious did you change the pin numbers on the part or did you simply move them?

You're coming in late and missing the point.  The result of my editing a symbol was the revelation
that schematic capture has no netlist.  The lack of a netlist in schematic capture is the point.
KiCAD only generates it later, as a postprocessing step from the completed schematic.

But it still seems Kicad has -by the looks of it- a complete electrical rule check in the schematics editor. So its not like the schematic editor is completely dumb.

Again: it would be nice if you could outline the benefits of having a netlist aware schematic editor in Kicad. Without getting sidetracked by 'Mentor does it this way' or Mentor specific user interface behaviour. IOW: what is the bare to the bones benefit to users to have a netlist aware schematic editor? What kind of useful features does it allow?

People developing Kicad are reading along here and I'm sure they are interested. But be prepared to have a long winded discussion because there are people that indeed think 'I don't need this so nobody needs this'. It has taken far too long for Kicad to support a database driven component library for example. For these kind of pro features to be added it takes a very clear and concise definition of the problem and possible solution in order to make the developers see the benefit and subsequently spend time on developing a feature.

[Monkeh]

there are people that indeed think 'I don't need this so nobody needs this'.

Bear in mind we're dealing with an equally extreme viewpoint, "I need this so everybody needs this".

[nctnico]

there are people that indeed think 'I don't need this so nobody needs this'.

Bear in mind we're dealing with an equally extreme viewpoint, "I need this so everybody needs this".

But still, it could turn out to be something worthwhile to persue. I'd like to have a clear understanding of what the problem is and how that can be solved.

So far I understand:
- Realtime ERC
- Being able to got through a list with nets in a schematic (and jump to the sheets where that net is being used)

[propellerhead]

Again: it would be nice if you could outline the benefits of having a netlist aware schematic editor in Kicad. Without getting sidetracked by 'Mentor does it this way' or Mentor specific user interface behaviour. IOW: what is the bare to the bones benefit to users to have a netlist aware schematic editor? What kind of useful features does it allow?

You'll find a reasonably concise and descriptive example in reply #173.  Yes, I refer to Mentor,
but that's just because it happens to be the system I'm most familiar with, so it's easiest for
me to walk through the steps.  There's absolutely nothing about it that is or should be specific
to that package, so if you don't like the word "Mentor", mentally substitute "Mumble".

One more time, though, I'll repeat the benefit:  It provides automatic, real-time error-checking
as you enter your schematic.  Not full ERC, but the most basic confirmation that what you see
on the page is really what you want, and makes it difficult to insert mistakes into it during the
course of simple operations like moves or symbol edits.  If the schematic keeps track of the
netlist as you're entering the drawing, it's able to spot inconsistencies so you can correct them
before they make their way to postprocessing.

[baldurn]

As others in the conversation have noted, there are examples of both methods
out there among the different packages, but in general, the higher-end, more mature, and
(incidentally) more expensive ones maintain and enforce the netlist in real time because
there's significant value in schematic capture understanding the difference between things
that are connected and things that aren't, and KiCAD has no such understanding.

I am wondering, do you have any background as a developer of large software packages? You come over as mighty arrogant and then you write stuff like "KiCAD has no such understanding" yet claim you are not talking about the underlying data structures - but that is exactly what you are talking about, you probably just lack the theoretical background to realize this.

KiCAD is a computer program and as such has no understanding about anything. But it can store data in various ways (in memory or in a file). It seems to store connections indexed by coordinates, which is indeed a bit silly. Instead it could store connections using identifiers. Or it could group them in what you think of as a "netlist", because of your background, but us software people would think of in more generic terms. There are infinitely many ways to do this. However only the end result actually matters and that could be achieved by fixing two bugs such that mr. propellerhead couldn't tell the difference.

In any case, the response is baffling, because they're arguing
for a tool that does less work for them and makes it easier for them to drive errors into their
designs, apparently because they think it's right because that's how it's always been done here.

No I am reading that most of us actually agree that KiCAD is doing something less than optimal here. Some of us may just not agree that it is such a big deal. It is not going to stop me from using the program and I don't have your option of using your old HP-UX workstation as an alternative anyway.

Since you don't hold back insulting anyone, I may be frank with you and say you come over as an old retired man that never got into the modern world. No cell phones, stuck using a computer and software older than some people here. And now you tried something new, didn't like it but felt a need to explain to the world it was not you, but everyone else at fault. Doesn't help that you may actually have a valid point about something not good in KiCAD. It is still a silly reason and likely not the truth about why you gave up on something new.

[Raven Luni]

I cant believe this thread is still going.  I think I posted a bit of a cringeworthy reply like 3 years ago.  Anyway - in response to the 'if you dont like it make your own' crowd I'm actually doing exactly that, started a few months ago.  I'm aiming for something that will cater especially to visually impaired people - that way its something positive in its own right instead of just another 'up yours' to something I don't like.

Not much to show yet but here's a test shot from my symbol editor:

[propellerhead]

Since you don't hold back insulting anyone, I may be frank with you and say you come over as an old retired man that never got into the modern world. No cell phones, stuck using a computer and software older than some people here. And now you tried something new, didn't like it but felt a need to explain to the world it was not you, but everyone else at fault. Doesn't help that you may actually have a valid point about something not good in KiCAD. It is still a silly reason and likely not the truth about why you gave up on something new.

Jesus.Mother.Fucking.Christ.  When are you people going to stop jabbering on about my
personality and make an effort to understand and stick to the conversation about a beneficial
improvement to schematic capture?

Since you're so desperate to know about me that you're building an elaborate mental construct
over it, here they are, in order:  Younger than you'll ever be; will use a cell phone when I have to
(i.e. very seldom), but object to obtuse UIs, service prices that are among the highest in the world
because there isn't enough actual competition and our government is incompetent at regulating
them, I don't want Apple or Google to dictate what software I run on a computer I own (or to have
to jailbreak it), and I don't like surveillance capitalism; so what, I probably have t-shirts older than
you, and (as I've previously explained but you're too imaginative to have bothered to read) the
penguix laptop that I set up for KiCAD is ready, willing, and waiting when I think the package has
schematic capture I can use and trust.  And don't call me a liar unless you've got evidence and
not just wild speculation that fits your own mental model rather than reality.  Now go find some
teenage actress to stalk, because you're creeping me out.

[propellerhead]

I cant believe this thread is still going.  I think I posted a bit of a cringeworthy reply like 3 years ago.  Anyway - in response to the 'if you dont like it make your own' crowd I'm actually doing exactly that, started a few months ago.  I'm aiming for something that will cater especially to visually impaired people - that way its something positive in its own right instead of just another 'up yours' to something I don't like.

That's pretty funny - I just bumped into a thread that had been dormant for more than a year
and relit/hijacked it for a UI-related issue.  Welcome back, and wade in at your peril, because
this swamp is full of reptiles gnawing at my ankles!

[PlainName]

As pointed out earlier in this exchange, propellerhead (under the user name "via") has already brought this up a few months ago on the Kicad forum. We can save ourselves time and aggravation, since that thread has fully explored the matter and ends with a very nice summary by moderator eelik:

https://forum.kicad.info/t/symbol-editing-and-the-netlist/39271/110

Coincidentally, he joined EEVBlog just when that thread got closed, and has posted nothing but this rant about the netlist.

@Propellerhead: again, what are you hoping to achieve here? If you didn't get satisfaction over there, what makes you think you will here, or are you just wanting somewhere to blow off?

[propellerhead]

Coincidentally, he joined EEVBlog just when that thread got closed, and has posted nothing but this rant about the netlist.

Unimportant coincidence, so don't read anything into it.  I've been lurking here for many
years but busy running my own unrelated electronics groups as well as engaging in a lot
of non-electronic activity.  So I assure you I'm not engaged in some whacked-out crusade -
at least, nothing like the fabulists and fantasists here who are deep into writing fanfiction
about me.

My present plan is to continue to ignore KiCAD until it gets over this problem, by whatever
means.

[John B]

I've been pretty happy proof reading and double checking my symbols before I start connecting them up. Even if you don't get the solution you wish, start by addressing the self inflicted aspect of this problem.

[propellerhead]

I've been pretty happy proof reading and double checking my symbols before I start connecting them up. Even if you don't get the solution you wish, start by addressing the self inflicted aspect of this problem.

Again, the symbols aren't the problem.  Returning to schematic capture from a trivial symbol
edit was the event that revealed the deficiency I've been discussing.  And I've heard the "well,
if you just make sure your symbols are perfect first, no problem!  Why would anyone who knows
what he's doing ever have to edit a symbol later?"  I don't think that's a sound view.  Increasingly
as symbols get bigger and more complex (as do the designs they're in), it's not only perfectly
reasonable, but probable, that symbols are going to need tweaking as the drawing takes shape
around them.  My argument is for schematic capture doing real-time checking in order to make
sure errors haven't crept in as a result of various editing operations.  I'm promoting the idea that
it's the software's job, not yours, to help protect you from "self inflicted" errors.

[nctnico]

Again: it would be nice if you could outline the benefits of having a netlist aware schematic editor in Kicad. Without getting sidetracked by 'Mentor does it this way' or Mentor specific user interface behaviour. IOW: what is the bare to the bones benefit to users to have a netlist aware schematic editor? What kind of useful features does it allow?

You'll find a reasonably concise and descriptive example in reply #173.  Yes, I refer to Mentor,
but that's just because it happens to be the system I'm most familiar with, so it's easiest for
me to walk through the steps.  There's absolutely nothing about it that is or should be specific
to that package, so if you don't like the word "Mentor", mentally substitute "Mumble".

One more time, though, I'll repeat the benefit:  It provides automatic, real-time error-checking
as you enter your schematic.  Not full ERC, but the most basic confirmation that what you see
on the page is really what you want, and makes it difficult to insert mistakes into it during the
course of simple operations like moves or symbol edits.  If the schematic keeps track of the
netlist as you're entering the drawing, it's able to spot inconsistencies so you can correct them
before they make their way to postprocessing.

Just to be clear (as I have no experience with Mentor at all so you can tell me anything): what you are after is that if you move a symbol the connections will rubber-band along without creating connections that shouldn't happen. If yes, then I think that is a usefull feature that can prevent costly mistakes.

I'm a long time Orcad user and I can tell Orcad doesn't do such checks so you have to be carefull when moving components in order to keep the connections the way you intend. The old DOS version was even worse. When I was an intern at an electronics company they had me create a PCB layout from a netlist. Once the PCB was assembled & being tested the electronics engineer started giving me a lot of flack for making a mistake in the layout. Having quite a bit of layout experience at that point already I was quite sure I didn't because the DRC check wasn't showing any shorts. Long story short, it turned out the engineer had moved a component and DOS-Orcad had screwed up. Instead of 2 lines crossing, both lines where taking a right turn and thus making the wrong connection. This was not visible on the schematic at all! The same circuit was repeated 16 times on the board as well so there was quite a bit of rework involved. Modern day versions of Orcad automatically place a dot in such situations indicating a connection between the lines (and put them into the same net). You could succesfully argue that this leaves checking the schematic for continued correctness with the user.

[propellerhead]

Just to be clear (as I have no experience with Mentor at all so you can tell me anything): what you are after is that if you move a symbol the connections will rubber-band along without creating connections that shouldn't happen. If yes, then I think that is a usefull feature that can prevent costly mistakes.

Yes, that's just one case.  Since when you connect a wire to a symbol pin, you're actually connecting
a wire to a symbol pin, schematic capture is aware of the connection, and the wires rubberband as
you move the symbol.  Often that movement will cause wires to intersect with other pins, wire vertices,
etc., which might have the appearance of being connected.  But since they're not (according to the
netlist and the fact that you didn't explicitly connect them), they get flagged as crossings for you
to clean up.  As I've explained a few times, the action that drew my attention to this problem in KiCAD
was doing the simplest of symbol edits - the library 555 is really odd in terms of which sides the various
pins are on - at very least, 2,6, and 7 are generally accepted as being on the left side and 3 on the right.
So after sorting it out in the symbol editor (which was nice to use), I returned to the schematic and
instead of finding the right wires rubberbanded to the right pins, the wires were now misconnected to
the pins I'd moved.  Is this a dealstopper when you're editing an 8-pin symbol?  Of course not.  But if
it's a TQ144 CPU or a PQFP208 FPGA or a 1000+ pin BGA, that can be a mess.

I'm a long time Orcad user and I can tell Orcad doesn't do such checks so you have to be carefull when moving components in order to keep the connections the way you intend. The old DOS version was even worse. When I was an intern at an electronics company they had me create a PCB layout from a netlist. Once the PCB was assembled & being tested the electronics engineer started giving me a lot of flack for making a mistake in the layout. Having quite a bit of layout experience at that point already I was quite sure I didn't because the DRC check wasn't showing any shorts. Long story short, it turned out the engineer had moved a component and DOS-Orcad had screwed up. Instead of 2 lines crossing, both lines where taking a right turn and thus making the wrong connection. This was not visible on the schematic at all! The same circuit was repeated 16 times on the board as well so there was quite a bit of rework involved. Modern day versions of Orcad automatically place a dot in such situations indicating a connection between the lines (and put them into the same net). You could succesfully argue that this leaves checking the schematic for continued correctness with the user.

Sounds like you have a pretty good handle on the kind of problem I'm talking about.  And I would
absolutely agree with you that adding the dot is exactly the wrong way to deal with it - assuming
that because two wires are crossing they should be connected, so quietly just taking care of that
for you?  What the hell were those people thinking?  The behaviour I'm describing would put a dot
there, alright, but a red one indicating:  Buddy, you've got something here to sort out.

[edit]

Worth adding that the system I'm talking about doesn't flag every crossing, of course.  That'd
be a makework PITA.  It just flags the ones that are more likely to be mistaken for valid connections,
such as a line crossing a pin, vertex landing on a pin, vertex (or end) touching a line, etc.

[PlainName]

Worth adding that the system I'm talking about doesn't flag every crossing, of course.  That'd
be a makework PITA.  It just flags the ones that are more likely to be mistaken for valid connections,
such as a line crossing a pin, vertex landing on a pin, vertex (or end) touching a line, etc.

Altium does that: drag or otherwise cross a pin with a net and they auto-connect, often leading to adjacent-pin shorts when trying to work close to a part. And Altium is a grown-up product. I've also used several products which make a great point of being able to plonk down something in the middle of a net (or, sometimes, net collection) and have them auto-connect.

[nctnico]

Worth adding that the system I'm talking about doesn't flag every crossing, of course.  That'd
be a makework PITA.  It just flags the ones that are more likely to be mistaken for valid connections,
such as a line crossing a pin, vertex landing on a pin, vertex (or end) touching a line, etc.

Altium does that: drag or otherwise cross a pin with a net and they auto-connect, often leading to adjacent-pin shorts when trying to work close to a part. And Altium is a grown-up product. I've also used several products which make a great point of being able to plonk down something in the middle of a net (or, sometimes, net collection) and have them auto-connect.

I don't think Altium and Orcad being 'grown up products' means that their implementation is somehow right or wrong. IMHO the idea that you have to make a connection between parts specifically instead of having the CAD package making assumptions (like Altium and Orcad do) has a lot of merit. I can certainly see the added value of having the CAD package ensuring the integrity of the connections during editing so the chances of errors due to moving / shuffling parts around and/or changing symbols are much smaller.

I don't think this is particulary hard to implement; keep a netlist with connections between devices / pin numbers and only allow changes to this netlist in 'connection mode'. Any other edit that invalidates the netlist, results in errors.

Thinking about it: a PCB package will yell at you when you try to make the wrong connection. But why does the schematic package allow you to do this without warning?

[PlainName]

I don't think Altium and Orcad being 'grown up products' means that their implementation is somehow right or wrong.

Indeed, but the premise seems to be that if they have this sort of behaviour then they are toys and not worth bothering with. Propellerhead says that this is the reason he won't use Kicad.

Thinking about it: a PCB package will yell at you when you try to make the wrong connection. But why does the schematic package allow you to do this without warning?

Because you're not making it up as you go along with the PCB? The schematic is where you create the netlist; the PCB is where you use that netlist. Thus you want to be able to make changes easily in the schematic but are very unlikely to do so on the PCB. Although, if you think you do want to then you can, but a warning is sensible because the probability is that you didn't meant to.

[woofy]

I also seem to recall that orcad was not aware of netlists. I've never used it, this is based on a visit to a client many moons ago. They designed theatre lighting control desks and used orcad. Their computer was unavailable on the day because they were "doing netlist extractions". This was back in DOS days. Not really relevant now but as orcad was mentioned.. .  .

I agree that kicad should have an inherent awareness of the netlist. but not for any of the reasons mentioned so far. My one real feature wish for kicad is to enable back propagation of changes to the netlist from PCB to schematic. i.e. back annotation or pin/gate swapping as some would say. I know not everyone thinks this is needed but its a feature I use all the time in Proteus with processor pins. When designing the symbol, I define groups of pins as swap-able. Then when laying out the PCB I can drag net lines from pin to pin to make the layout easier (it swaps the nets on the to and from pins). These changes back propagate to the schematic, but not by moving any wires. That would lead to the kind of schematic mess already described. Instead it swaps the pins on the schematic symbol. That way a neat schematic layout is maintained.

However I don't see how such a feature could work without the schematic having an intrinsic knowledge of the netlist. That would seem to be a necessary first step.

[nctnico]

I don't think Altium and Orcad being 'grown up products' means that their implementation is somehow right or wrong.

Indeed, but the premise seems to be that if they have this sort of behaviour then they are toys and not worth bothering with. Propellerhead says that this is the reason he won't use Kicad.

Everybody is free to set their own standards. IMHO it is worth debating the merits of certain features as it can only lead to improving Kicad and make it useful for a larger audience.

For example: personally I won't use a CAD package that doesn't have a database driven component system. After years of debate and strong pushback from the 'I don't need it, so nobody needs it' crowd, a database driven component management system is finally implemented in Kicad.

Thinking about it: a PCB package will yell at you when you try to make the wrong connection. But why does the schematic package allow you to do this without warning?

Because you're not making it up as you go along with the PCB? The schematic is where you create the netlist; the PCB is where you use that netlist. Thus you want to be able to make changes easily in the schematic but are very unlikely to do so on the PCB.

Yes, but what if the change to the schematic is unintentional (when using an editing mode that isn't intended to create/change connections)?

[PlainName]

but what if the change to the schematic is unintentional

You get into the realms of mode switching. I don't think you need a pukka netlist if modes are used, but I haven't written one of these things. Maybe ask Nigel

Actually, that's something that annoys me with Altium. In most modern interfaces you select something and then command something to happen to it. Let's say changing the z-order of a window in on Windows: select the window, right click the system menu (both of these could be one operation, of course), select 'stay on top' or whatever your window manager calls it. In Altium to decide what you're going to do and then do it to objects. So for this example you would open some menu (of course, not a right click context menu but some global one) and select 'stay on top', then click the window to which you want it to happen.

I understand how it go to be like that, and that it can even be a Good Thing if you're doing an op on lots of things. Plus it's a mode, so you can't accidentally, say, roll up a window by mistake in that example. But I find it a drag because often I use that particular thing just to make some filled box be in the background when making a component. I think it's quite rare that I am doing the same op to many objects, and if I would it wouldn't be that much aggro to preselect them all anyway.

[tooki]

but what if the change to the schematic is unintentional

You get into the realms of mode switching. I don't think you need a pukka netlist if modes are used, but I haven't written one of these things. Maybe ask Nigel

Actually, that's something that annoys me with Altium. In most modern interfaces you select something and then command something to happen to it. Let's say changing the z-order of a window in on Windows: select the window, right click the system menu (both of these could be one operation, of course), select 'stay on top' or whatever your window manager calls it. In Altium to decide what you're going to do and then do it to objects. So for this example you would open some menu (of course, not a right click context menu but some global one) and select 'stay on top', then click the window to which you want it to happen.

I understand how it go to be like that, and that it can even be a Good Thing if you're doing an op on lots of things. Plus it's a mode, so you can't accidentally, say, roll up a window by mistake in that example. But I find it a drag because often I use that particular thing just to make some filled box be in the background when making a component. I think it's quite rare that I am doing the same op to many objects, and if I would it wouldn't be that much aggro to preselect them all anyway.

While I agree with you about the advantage of thorough consistency of the predominant “noun-verb” paradigm with Mac and Windows, that’s actually what Altium uses almost everywhere. The z-order commands (send to back, etc) are some of the exceptions that use “verb-noun” instead, as are some of the room management commands.

[PlainName]

Yes, and probably another reason why they are so irritating!