Kicad - GUI is Horrific!

[nctnico]

Going a bit off-topic here but out of interest:
@Eugene: since you seem to have experience with both Kicad and Altium: what is your assessment on how far KiCad is from becoming an actual replacement for Altium? My feeling is that the next version of Kicad could hit that mark or am I very wrong?

[twospoons]

[

Now you know of two professional EE's that use and like KiCad. You are no longer entitled to say "I have never met a person in a professional company liking KiCad." 

Make that three.

My story is remarkably similar to eugene's.  Spent most of my career using PADs and Altium, but when I went contracting I chose Kicad because I couldn't justify the cost of an Altium license, and Kicad was capable of doing what I want.

[Bassman59]

[Childish and passive aggressive response deleted.]

I'm frustrated because this thread is five pages and four years of going in circles. The complaints seem largely unchanged even though the thing being complained about has changed. Moreover, those changes have been a direct result of the complaints, yet the complaints keep coming more or less unchanged.

I used to participate in a forum dedicated to professional-level audio electronics design. For whatever reason^* a couple of years ago I stopped participating, but recently they updated the forum software and I started getting daily post digest emails.

One of the posts was an old thread complaining about Kicad's user interface -- basically a clone of this thread. Everything you read here was there: "Programmers are volunteers and only work on what interests them" and such.

And yes, there was a handful of posts from someone complaining about the user interface, and I wrote a post saying, "your complaints are obsolete -- everything you're complaining about has been fixed, and in cases X and Y they were fixed two years ago."

That got a response claiming that "Kicad users are zealots!" My response to that was basically, "Wait, so you ignore that your complaints were actually addressed, even though you've never filed a bug-fix or wishlist request, and you refuse to give the latest version a try to verify this, and you have the nerve to call me a zealot? I'm out. Bye!"

I would delete my account and posts there if I could. The really silly thing is that the complainer is a forum moderator.

Anyway. This thread appears in pretty much all electronics forums.

^*After looking at recent threads after getting those digest emails, I realized that the reason I stopped participating was because the forum had become a circle jerk. The same people talking endlessly about the same topics. There were no new-user posts.

[eugene]

Going a bit off-topic here but out of interest:
@Eugene: since you seem to have experience with both Kicad and Altium: what is your assessment on how far KiCad is from becoming an actual replacement for Altium? My feeling is that the next version of Kicad could hit that mark or am I very wrong?

I think you're right. Obviously it depends on what features of Altium you depend on. For me, there's nothing that I've wanted to do but couldn't get done with KiCad since I started using it in 2018 (early V5.) Would it be a good idea for a large tech company to dump Altium and have everyone switch to KiCad? That might be problematic, but I believe it could be made to work today with V6. The basic functionality is there. As I wrote in my last post, all of the well-known tools, no matter how expensive, have issues.

There's no doubt that Altium has a more mature interface, but I'm not sure that actually makes it better. 'Mature' eventually turns into 'old.'   

The lack of an automated parts database might be the biggest issue, but I know that only from the complaints of others. It is not an issue for me. In any case, it seems that the problem is already solved, or will be soon, through plugins.

[james_s]

Anyway. This thread appears in pretty much all electronics forums.

^*After looking at recent threads after getting those digest emails, I realized that the reason I stopped participating was because the forum had become a circle jerk. The same people talking endlessly about the same topics. There were no new-user posts.

People like to complain, it feels good to complain, and some people do more complaining than others. I don't really see it as a problem unless it does what you mention there and completely takes over the forum. This forum on the other hand has a wide variety of content and new threads every day.

[Bassman59]

There's no doubt that Altium has a more mature interface, but I'm not sure that actually makes it better. 'Mature' eventually turns into 'old.'   

For some reason, the rocket surgeons at Altium decided to take the schematic/layout cross-probe button off of the main icon bar. Idiots.

[John B]

Has anyone noticed that most electrolytic cap footprints no longer have positive terminal markers on the silkscreen layer?

edit: Nevermind....I realised there's a whole host of non polarised footprints too now.

[PlainName]

I have no axe to grind since I can't use Kicad now, but I've previously spoken about my frustration with the interface.

The complaints seem largely unchanged even though the thing being complained about has changed. Moreover, those changes have been a direct result of the complaints, yet the complaints keep coming more or less unchanged.

Perhaps the changes didn't fix what was being complained about. As an example (and nothing to do with Kicad, just some random example), maybe a bedside clock has a red LED flashing every second, and people complain about that. So the developer changed the LED to a blue one - voila, problem fixed. Or not

[b_force]

But anyway, not that it matters, because it seems 99% impossible to have a constructive and respectful conversation about it.

I wonder how many Altium or Eagle/Fusion software developers design PCBs for a living?

This kind of snappy responses do not help, and just add weight on b_force's statement that it is really hard to have a conversation regarding this.

In fact it's the main reason why I don't bother much to be on forums and these kind of discussions anymore.

These days I can even literally predict the next response; that they beg to differ, but totally missing the bigger picture.

Also instead of having an objective constructive debate, people just come up with personal believes that are already set and stone.
That is not a debate or an argument, that's just throwing statements at each other.
In a proper discussion or argument people try to understand each other points of view.

A company like KiCad who is selling a PCB program I would HIGHLY recommend listening to people with MANY decades of experience.
And YES, even if you try to convince people to use your FREE program, that IS still selling. (you only don't get money for it)

We will see were KiCad end in 5 years, like I did 5 years ago, my expectations are very low.
Up till now, my prediction were very accurate, but hopefully I am wrong this time.

[andrejr]

I'm a longtime KiCad user, and I've grown accustomed to it over the years. I'm pretty fast and efficient, and I use both hands (left one spams keyboard shortcuts, the right one moves the components).

But in spite of that, I find some workflow choices pretty dumb, unintuitive and plain bad. Selecton behavior comes to mind, some aspects of rotation, ... Now, I'm a C++ developer and I wanted to help make it better, but it uses WxWidgets which I have never seriously used. I'd have to invest time into learning about WxWidgets first, which puts me off of the whole thing (when I do GUIs, it's Qt).

The simple file formats, plugin functionality and Python integration in KiCAD make it so much better than any other solution available for any amount of money. You can't reproduce productivity offered by KiPart, KiField, Kinjector, KiCost, and the ability to write your own scripts. Then there's integration with Inventree, the FOSS inventory app (it's called Ki-nTree and I'm a contributor). The amount of symbols and footprints for obscure components is also unmatched.

If you're the oldschool EE who doesn't know much about programming (scripting), but needs an EDA package to do schematics and boards, you might be better off with some other tool. If you're doing really difficult boards (GHz range signals, huge BGAs, ...), I'm also not sure it's the right choice (though there is some script-based auto-fanout support) because I don't do those. I have no idea how it handles microwave stuff (though I know it has some tools for it). But if you're doing anything else, it's the best choice.

In spite of that, there is a lot of room for improvement, in both the GUI and component management.

[nctnico]

But anyway, not that it matters, because it seems 99% impossible to have a constructive and respectful conversation about it.

I wonder how many Altium or Eagle/Fusion software developers design PCBs for a living?

This kind of snappy responses do not help, and just add weight on b_force's statement that it is really hard to have a conversation regarding this.

In fact it's the main reason why I don't bother much to be on forums and these kind of discussions anymore.

These days I can even literally predict the next response; that they beg to differ, but totally missing the bigger picture.

What is the bigger picture? That by itself is a bit of an empty statement. For sure there are other more advanced CAD packages out there compared to Kicad. But there are also less advanced ones. The way I see it KiCad is somewhere on the middle of the road and has replaced Eagle as a general purpose schematics & layout tool.

From my interactions with other EEs I know there is a wild variety of workflows that people use. Some even prefer to use a glorified paint so they have full control over where their traces go. Things like netlists and DRC just get in their way. At the other end of the spectrum there are EEs that like to use all kind of advanced features like area based DRC, BGA fanout, differential pair constraints, etc. There is a very large area in between as well. In the end better is relative to what functionality you need.

Where it comes to the GUI: in my experience every piece of software that does a somewhat complex task, has a horrible GUI that simply needs time to learn.

BTW: I'm not a Kicad user at all. I'm just interested to see where it goes and maybe I'll hop on at some point depending on the circumstances.

[ebastler]

In fact it's the main reason why I don't bother much to be on forums and these kind of discussions anymore.

These days I can even literally predict the next response; that they beg to differ, but totally missing the bigger picture.

Also instead of having an objective constructive debate, people just come up with personal believes that are already set and stone.
That is not a debate or an argument, that's just throwing statements at each other.
In a proper discussion or argument people try to understand each other points of view.

If you want a constructive discussion, you might start by being a bit self-critical about your earlier post. Do you really believe that, for Kicad to have merit, its programmers also must be professional PCB designers? Do you really believe that Kicad has not improved at all over the years, as you had implied?

You just throw out these polemic statements and, at the same time, complain about a lack of constructive argument. Notice something?

The good thing is that we don't need to have an argument at all. Since in your view Kicad is beyond hope, I am sure you are using a different PCB CAD package, and I certainly don't want to get in the way of that. While you do so, just let those who are content with Kicad use it, and let them discuss among themselves which improvements they consider most relevant.

[b_force]

If you want a constructive discussion, you might start by being a bit self-critical about your earlier post. Do you really believe that, for Kicad to have merit, its programmers also must be professional PCB designers? Do you really believe that Kicad has not improved at all over the years, as you had implied?

And this is the exact reason why I don't bother.

Proper reading and thinking in black/white statements will always results in arguments that don't go anywhere.
Forever talking in circles by already assuming things (as well as having those statements ready)
Besides that, you are missing the point of the bigger picture, the point is not if something has improved.
The point is the general overall vision of something.

You can improve a turd, it might look pretty and shiny, but it always will be a turd.

Besides I don't really care tbh, but I ques that's considered as rude?
Can never win, can we?

I will just have a look in a while again, cheers! 

(waiting again for the responses; "yes, but it's free", "oh, but can you do it better", "Have you ever made PCB program", "Works fine to me", "Have you ever done PCB design" etc etc etc
https://en.wikipedia.org/wiki/Fallacy)

[thinkfat]

IMHO, KiCAD is not going to replace Altium in the foreseeable future. V6 was not as big a step forward than V5 was against V4. For me, V5 was plenty enough and V6 felt like just an incremental improvement. V5 was a major leap forward, in comparison. There's nothing in V6 that enables me to do things that weren't possible already with V5. Had they brought the simulation engine up to par with LTspice, or implemented a seamless interface to e.g. openEMS, that'd have been "wow" stuff.

Still, KiCAD is by and large a stand-alone tool. It does what it does quite well but for use in a big production environment, it just lacks connectivity towards the dreadful world of ERP systems, that every big company apparently must have. And since this stuff is just boring and unattractive, nobody will want to work on that in KiCAD unless forced with money. Changes in the user interface are not going to make a difference, even if it was the spitting image of Altium Designer, it would still not replace it.

[aaronw]

I can't speak to other platforms, but a few years ago I decided to try my hand at making a PCB to do what I wanted. After struggling with the schematic editor for Eagle I gave up in frustration and downloaded KiCad. I found KiCad to be far more intuitive to use. Right-click popped up context menus and the keys were generally intuitive. Want to delete a component? Select it and hit delete. I've used it ever since. There's certainly room for improvement, especially with all the UI bugs I'm hitting with 6.0. I'm hitting a serious problem where the focus switches away from the PCB editor to the schematic, so when I select a component to rotate with 'R' I might suddenly have the same component in the schematic editor rotate! This may be some weird interaction with KWin and the fact I use focus follows mouse. I'm sure that will be fixed since I'm not alone with this.
I think part of KiCad's problem was their choice of wxWidgets compared to something like Qt, the latter being better for large complex applications with custom GUI requirements. I also hit a lot of issues with routing in 6.0 where it seems to struggle at times, and clicking a track does not lay it down to let me continue, much to my chagrin. In many ways, I prefer 5.1. It certainly worked a lot better for me.

[MadScientist]

High end Altium users are not in the sights of KiCad , but certainly professional Eagle and low end professional Altium users are.

Hence it wrong to preset this as an all or nothing argument.

Note V6 was a huge improvement  in that it bundled the libraries of parts used into the schematic directly. This now means exporting the project to others is more less error prone.

[MadScientist]

I can't speak to other platforms, but a few years ago I decided to try my hand at making a PCB to do what I wanted. After struggling with the schematic editor for Eagle I gave up in frustration and downloaded KiCad. I found KiCad to be far more intuitive to use. Right-click popped up context menus and the keys were generally intuitive. Want to delete a component? Select it and hit delete. I've used it ever since. There's certainly room for improvement, especially with all the UI bugs I'm hitting with 6.0. I'm hitting a serious problem where the focus switches away from the PCB editor to the schematic, so when I select a component to rotate with 'R' I might suddenly have the same component in the schematic editor rotate! This may be some weird interaction with KWin and the fact I use focus follows mouse. I'm sure that will be fixed since I'm not alone with this.
I think part of KiCad's problem was their choice of wxWidgets compared to something like Qt, the latter being better for large complex applications with custom GUI requirements. I also hit a lot of issues with routing in 6.0 where it seems to struggle at times, and clicking a track does not lay it down to let me continue, much to my chagrin. In many ways, I prefer 5.1. It certainly worked a lot better for me.

I’m on V6 and I’ve done about 10 designs with it. Ive never experienced the focus issue you outline. You need to examine your workflow. I use a dual monitor arrangement

[aaronw]

I'm using dual monitors in Linux with KDE and I have my desktop environment configured to focus follows mouse. What happens is in pcbnew, sometimes when I select a component, the focus will shift to the same component in eeschema and the window focus will shift from pcbnew to eeschema so that if I then type a key, like 'R' or 'M' it will affect the component in eeschema instead of pcbnew. I opened a bug on the KiCad site with instructions on how to reproduce this reliably. I never had this issue with 5.1.x. The other issue I have is how laying down tracks has changed. With 5.1.x if I click the mouse button it will place the current track where it sits and let me continue. With 6.0 it doesn't do this and it will continue to try and push and shove things around unless I click at least twice, where it also will often just end the track where it is. And finally, it seems to have a lot more trouble hunting around for placing a track than 5.1 did.

Note that my PCBs tend to be rather dense. https://github.com/aaronw2/led-controller-v4 is my current work in progress.

[propellerhead]

No one is forcing you to use KiCad, so use Circuitmaker?  I don't see the problem.
It's funny how entitled people get when you give them free stuff.

Yes, it's been a year since this thread cooled, but I think it's time for it to reheat.

What we have here is an argument for the basic acceptability of mediocrity, in this case,
"It doesn't have to be good if you're not paying for it."  An impressive position.

I'm a longtime user of really grownup EDA; which one doesn't matter, but I don't mind mentioning
if it's relevant to the discussion.  The point is that it's also quite an old package, and running on a
Unix workstation with a big glass tube.  So as much as I love it, I sure could use a little modernization
onto, like, something that runs on a laptop, but without windoze because anything newer than xp
has been insufferable.  Anyway...

That's what led me to give KiCAD a go at the suggestion of a friend whose recommendation I felt
should be reliable.  As a test project I set out to recreate a really old through-hole design, a couple
of boards that although very few people on earth have a need, the ones who need it really need
it.  Technically, unchallenging in the extreme.  Little did I anticipate the nightmare that awaited.

I was doing my first pass, entering the schematic from an almost-unreadable copy of a hand-drawn
original, and found that the stock library 555 symbol was really wonky, completely out of keeping
with the way it's normally done, inputs on the left, output on the right.  So I crack the symbol editor
and find they have some nice touches, like giving you the choice of adding an inverting circle, which
is something I'm used to having to manually draw in.  Then I exit back to schematic capture and find
that the connections I'd made before the edit didn't stick with the pins and rubberband; instead, the
moved pins are sitting on the wrong wire ends, which haven't moved.  Editing a symbol trashed the
netlist.  Not cool, not even a little bit.  Must have done something wrong, missed an update step or
something.  Took the question to the forum and was treated like a heretic to be placed in stocks in
the public square and stoned to death.

The part I quickly found incomprehensible is that this isn't a bug, it's a feature.  When you think you're
connecting a wire to a pin, you're not.  All you're doing is putting an end of a line at a place where there
happens to be a dot.  KiCAD's schematic capture does not join the two and add the connection to the
netlist database, because there isn't one.  Somewhere later in the design flow, I guess, some netlisting
module must run around and generate meaningful data, but at that point you're just drawing a bunch
of extremely error-prone pictures, because any time you move something, lines and pins and vertices
are going to bump into each other, and without a running netlist db to verify against, KiCAD can't tell
you when you're about to screw yourself.

Over the course of a hundred exchanges, I heard arguments that most of the friends I've talked
about it with (professional PCB designers and not) have a hard time believing had been seriously
offered:

The first, most obvious, and most condescendingly passive-aggressive:  "That's just your opinion. 
There's no 'right way' to do this.  You'd probably be happier using something else."

"Why would anyone have to edit a symbol?  Do it right the first time and you'll have no reason to
edit it later."

"Schematics exist EXCLUSIVELY as human-readable documentation.  They're not supposed to build
a database or anything else."

From the friend who recommended it, but did not participate in this bun fight:  "The project was
started by a guy at CERN!", implying that being associated with a multibillion-dollar Big Science
project confers the Trademark of Quality on some guy in someone's department's software project,
and ignoring the fact that giant organizations like that have budgets discretionary enough to let
just about anyone take a stab at some new crap if the argument is halfway good (e.g. It's for an
open-source package that'll sweep the world!), and that they also have more than their fair share
of boobs on staff.

The litany of idiotic arguments really is pretty stunning, go read it yourself if you must.  The
short answer is that my personal, professional determination is that unless and until KiCAD
reconsiders their shockingly naive and trivial approach to schematic capture, it's Greasy
Kid Stuff and not ready for prime time.  Maybe their layout is brilliant, but I'll have no
chance to find out as long as schematic is crap.

The other point this underscores is that open-source software users exhibit cult victim behaviour.
Not universally, there's a lot of both good and bad, but there's always tribalism.  And there's
certainly the idea that because open-source is morally superior to commercial software
(because it's being done for righteous reasons, and not just filthy profit), that moral superiority
somehow confers technical superiority - and that's where they go off the rails, sacrificing
any reasonable degree of objective judgement.

And... oh, yeah... the GUI's really awful too.  These people (the developers) have obviously never
heard the old saying about how great composers don't invent - they steal.  Go and nick the best
ideas available for your GUI from every piece of software you can get access to and roll them
together into a better one.  NIH is for idiots.

[wilfred]

Well you've covered a broad series of grievances I think that should help advance the development of Kicad. I'm surprised after a hundred exchanges with others you still wish to persist with Kicad. Surely trying something else might yield better results for you.

[propellerhead]

You must have lost count, because the only "grievance" I've discussed is schematic capture's
designed-in ignorance of the netlist.  That's all.  And it's not even a grievance.  I'm warning other
people who might mistake it for a serious tool that by any experienced, objective measure, has
failed in one of its most basic design requirements.  I'm no more reporting a "grievance" than is a
traffic reporter calling in a car accident.  I'm just passing on the facts and couldn't care less if
anyone wants to settle for such.

You're also making an incorrect inference if you think I "wish[ed] to persist".  I didn't.  I ditched that
crap as fast as I could and have happily returned to the system I've been using for something over
20 years, and I'm prepared to keep running it as long as it's light-years ahead of whatever open-
source stuff the kids are smugly cranking out now.  A system designed by people who understand
that the "A" in "CAD" means "Assisted", a tool that helps me drive through an error-free design by
catching problems rather than ignoring them because it can't tell the difference.

And apparently you skimmed my post too quickly to notice that I'd already addressed in advance
your ever-so-helpful "Surely trying something else might yield better results for you."  Yup.  Trying
something that I already know isn't garbage yields fabulous results.

[KE5FX]

Well you've covered a broad series of grievances I think that should help advance the development of Kicad. I'm surprised after a hundred exchanges with others you still wish to persist with Kicad. Surely trying something else might yield better results for you.

He's got some good points.  One thing I actually appreciate about EAGLE is that the schematic is the controlling document, so to speak. From what he's saying, it sounds like KiCAD not only still doesn't do back annotation properly, in the sense that coherence between the board layout and schematic are continuously enforced, but that it treats connections as "artwork" rather than as first-class objects. 

I can see where a lot of people would find that approach problematic while others might not care as much.

[wilfred]

He's got some good points.  One thing I actually appreciate about EAGLE is that the schematic is the controlling document, so to speak. From what he's saying, it sounds like KiCAD not only still doesn't do back annotation properly, in the sense that coherence between the board layout and schematic are continuously enforced, but that it treats connections as "artwork" rather than as first-class objects. 

I can see where a lot of people would find that approach problematic while others might not care as much.

Kicad does have to continue evolving to get out from under the limitations of its roots when it was a mash-up of tools by different developers. That's happening faster now than previously. I've seen an increasing number of people using it to good effect.

In any case my post was seeking an answer to an underlying question. Which has been answered I see.

[propellerhead]

Yeah, the forward/back annotation thing can be a double-edged sword.  Back when I
got started using PC-based EDA in the mid-80s (after almost 10 years of tape and mylar),
I went with a fully-integrated package (EE Designer) rather than the ad-hoc stuff friends
were using, usually OrCAD and whatever for layout.  That gave me an early appreciation
for being able to work in either direction with a minimum of fuss and/or error.  Being able
to draw a layout and then extract the schematic from it was particularly helpful when
doing reverse-engineering jobs.

The Tough-Guy system I've been using since making the jump from PCs to workstations
doesn't afford that luxury:  The schematic is the word of god, and if you don't go back to
the source drawing and explicitly make a change, you ain't making it in layout, because
that's introducing errors.

To clarify:  I have no idea whether it back-annotates (properly or not) from layout to
schematic, because I never got that far.  I wasn't about to waste a New York minute with
a system that doesn't know what its basic job is, so I have no idea what layout is like. 
And if it does back-annotate correctly, what's the point if the schematic capture still
doesn't understand what a netlist is?  Honestly, when I think about it, it's the software
equivalent of Monty Python's Swamp Castle.

[thinkfat]

The part I quickly found incomprehensible is that this isn't a bug, it's a feature.  When you think you're
connecting a wire to a pin, you're not.  All you're doing is putting an end of a line at a place where there
happens to be a dot.  KiCAD's schematic capture does not join the two and add the connection to the
netlist database, because there isn't one.  Somewhere later in the design flow, I guess, some netlisting
module must run around and generate meaningful data, but at that point you're just drawing a bunch
of extremely error-prone pictures, because any time you move something, lines and pins and vertices
are going to bump into each other, and without a running netlist db to verify against, KiCAD can't tell
you when you're about to screw yourself.

[...]

And... oh, yeah... the GUI's really awful too.  These people (the developers) have obviously never
heard the old saying about how great composers don't invent - they steal.  Go and nick the best
ideas available for your GUI from every piece of software you can get access to and roll them
together into a better one.  NIH is for idiots.

One of us must be missing something. What exactly do you mean with the schematic capture not joining pins and wires and adding that information to the netlist? KiCAD quite obviously "knows" in the schematic editor, that wires are nets and what they connect to, and it becomes abundantly obvious as soon as you hit the ERC button.

Regarding the GUI, beauty is in the eye of the beholder. I can find no (major) fault with KiCads way of interacting with me, but that may just be my personal bias (or lack thereof). As far as "steal is better than invent", I'm sure glad the developers didn't copy from "Eagle".