Alternatice UI

[delfinom]

> Why? Why not just let the thing be dragged with a left mouse button drag?

Ok.
So I'm trying to understand your description which I can now read better because claiming this is "Windows behavior" made no sense at all.
Are you saying you want to click on items and while still holding the mouse button be able to immediately drag it?

dunkemhigh's earlier description seems very clear to me. It's how things work in any standard Windows application, and as far as I am aware also on MacOS and (most?) Linux GUIs:

My suggestion is to copy any Microsoft Windows app: click-drag to select, leave selected but free the mouse. Only if the left button is pressed inside the selection should the selection be dragged or moved.

I've used Windows since Windows 95. I've never considered this a "Windows thing".
I am somewhat interested in implementing it behind a setting flag, but there's only a week left until feature freeze for V6 so it might not happen in time :/

[PlainName]

Thanks for the suggestion of DipTrace. However, I am not looking for a free or 'hobbiest' product per se - I would expect to use it for professional work as well as personal designs, and I don't object to paying for a good tool. I am also not short of PCB software already.

My reason for looking at KiCAD was because I am interested in what's around and what features they have that might benefit (or just interest) me. It gets mostly good reviews and looks like I might need it as a skill in a professional capacity at some point, so I thought it worth spending a little time using it to knock off an actual design (the best way to get into a product, I find).

[Doctorandus_P]

Some 6 or 7 years ago I was in need for a new PCB suite, as Windows tried to force a "Blue tiles of Death" screen upon me on every boot, and that was the final straw that made me switch to Linux. So I tried a few programs. Eagle had a completely horrible user interface. I could not even type in a value, and the first schematic I made in it had a 0.2Ohm resistor, that value was not in the drop down list.

Also tried Geda. It was usable, sort of, but the schematic and PCB part were completely separate programs with different sort of interfaces.

When I tried KiCad, together with the "Getting started in KiCad" guide (which is obsolete now unfortunately) I got a pretty good idea of how the user interface worked and it worked quite well. Back then there were still a handful of serious bugs in KiCad. Bugs on the level of: you can edit a library component, but then you can't save your edited component.

Then in the last 5 or so years I've seen KiCad grow. All serious bugs got smashed, new functionality added, GUI improvements, tremendous libraries. Side projects to automate boring tasks, a friendly forum...

The click-drag-select-move you mentioned is currently only present in Eeschema. In Pcbnew the selection is already separate from the operation, and it has been so for quite some time. Eeschema is getting a mayor update with KiCad V6 (Expected within a year probably) and the idea is to get it more in line with Pcbnew.

For the still limited amount of developers and budget for the KiCad project they've done an amazing job in these last years.

Maybe I'll let them have another go at it before jumping in to show 'em how it's done

Go on, jump in. Show them how to code a better GUI.

[Doctorandus_P]

No, I am disappointed with being in a mode. What tends to happen is you go to some mode (say, placing a track) and click, click, oh just check the incoming email/reference pdf/EEVblog so send your mouse to a different window and it doesn't go because it's locked into track placement mode inside that window. And you've just panned 20 feet away from what you were placing because the window pans when you hit the edge. Etc.

If you want some productivity, then get rid of those distractions. Quit your mail program, close internet, find some focus to work on a project.

I also do not like the panning, and turned it off long ago. Setting is:
**Pcbnew / Preferences / Preferences / Common / Pan while moving object**

I do have the "Center and warp cursor on zoom" checkbox on (default). This is a feature lots of users bitch about, (you probably too), but if you give yourself some time to get used to it, then you start wondering why other programs do not work in that way.

[PlainName]

The click-drag-select-move you mentioned is currently only present in Eeschema

Is Eeschema a different product to Kicad? Come to that, is PCBNew? (And what does PCBOld do?) When I download and install Kicad, what am I actually installing if not those?

    Maybe I'll let them have another go at it before jumping in to show 'em how it's done

Go on, jump in. Show them how to code a better GUI.

Hmmm. Maybe I shouldn't try to inject humour into things. Clearly, even adding a smiley emoticon doesn't show intent nowadays.

If you want some productivity, then get rid of those distractions.

Those were generic examples anyone could identify with. But real-life ones might be browsing to a file, opening a datasheet, running a calculator, checking JLCPCB for limitations, looking up footprint dimensions, ... Do you really need me to list a ton of stuff to which you can try your best to imply the user is at fault for not devoting 100% of their - and their machine's - time and focus to this single app? It is the app that is failing to follow accepted user interface guidelines, remember, not the user.

[Doctorandus_P]

When you install KiCad, it's made up of about 10+ different programs probably:
* KiCad project manager (normally started for each project)
* Eeschema, this is the program for the schematics.
* Pcbnew, For drawing the PCB.
* Symbol Editor.
* Footprint Editor.
* Footprint Library manager.
* CvPcb for assinging footprints to schematic components (called differently now-a-day's)
* Gerbview for viewing gerber files.
* Various utilites, such as "Bitmap to Component", the "Page Layout Editor" and more.
* Footprint scripts (in Python!) BOM scripts ...

And all of these work reasonably together to form the KiCad EDA Suite.

KiCad is pretty much "Modeless" Drawing a long track in KiCad is comparable with drawing a polyline in almost any CAD program. Having to go through a 3 layer deep menu structure for each line segment would probably not be a GUI improvement. So what's left is the panning off screen, which is easy to turn off.

[LazyJack]

Regarding the drag mode. In most cad software, you usually don't just want to drag something somewhere, but specify exactly the drag. You may want to select the reference for dragging. In a cad tool you can select a bunch of objects, then say move, then say you wan't to move the centerpoint of a circular part to the intersection of two lines. This is much harder to do by just clicking and dragging, but easy to do with drag mode. You may want to be in a completly different zoom setting and part of the drawing when selecting the objects, selecting the reference and when dragging.
This is why the majority of cad tools work this way.

[PlainName]

When you install KiCad, it's made up of about 10+ different programs probably:

Yes, I realise that. My question was rhetorical ("asked in order to produce an effect or to make a statement rather than to elicit information" - ODE). To expande, if you (the generic 'you', not necessarily you personally) complain that the forum doesn't embed images properly, the implication is that it's a forum problem. If you're a techy and privy to what runs the forum, you might be a pedant and say it's SMF. It's kind of stretching things to say it's such-and-such plugin problem, or even that it's a foible of php. To the user, it's EEVBlog. It's a coherent whole (or should be).

[PlainName]

In most cad software...

Oh boy! That's a messy can of worms to tip over

I like the way Altium implements drag selection: drag down and to the right selects anything completely enclosed, while dragging up and to the left selects anything with a part enclosed. But for other stuff they are very poor. Usually, one would select a bunch of things and then do something with them. With Altium you first say what you're going to do to them, then you select what you want to do it to. The result is that it's a bit longer to do anything (and if you're from a Windows world, which you will be, it's counter-intuitive).

This is why the majority of cad tools work this way.

But some seem to manage just fine. TurboCAD, for instance, doesn't seem to have a selection or dragging problem. At least, I can't recall being annoyed by it in those respects. In fact, now I think about it, it does pretty well: select stuff then do something to 'em, OR select what you want to do and then do it to things individually. But click'n'drag works just like you'd expect it to do. Well, like I'd expect it to!

If that app - a CAD tool no less! - can manage it, why can't others? The answer is probably because they have a long history and retain certain ways of doing things because that's how they were done. And, of course, they're all different because they all come from a time when there weren't any standards, or they we too difficult to implement. Or, again quite probably, because they are just a massive bunch of addons and fixes and enhancements to the first quite simple, and not really designed as such, utility.

Proteus is a good example, having its roots in DOS and dragging that user interface into Windows with it, but over time (25 years or so) they changed to a more Windows-centric interface. Still not quite correct, but a lot better than it used to be.

[Doctorandus_P]

In most cad software...

Oh boy! That's a messy can of worms to tip over

It sure is, but for some things you have to find a compromise somewhere.

I like the way Altium implements drag selection: drag down and to the right selects anything completely enclosed, while dragging up and to the left selects anything with a part enclosed.

Same in KiCad. In Pcbnew (that is the PCB part of KiCad ) it already works this way for quite some time. Eeschema (The schematic part of KiCad ) is likely to follow with the next mayor update.
Small difference: KiCad only distinguishes with dragging left to right, or right to left. It don't care about up and down, that is up to you.

Proteus is a good example, having its roots in DOS and dragging that user interface into Windows with it, but over time (25 years or so) they changed to a more Windows-centric interface. Still not quite correct, but a lot better than it used to be.

KiCad ain't even 25 year old! It started as a hobby project from a university professor, and it's only really taking off in the last handful of years. Before that it was pretty buggy and development was much slower then now. KiCad is picking up development speed each year. Give KiCad a few more years (it won't need 25) and then see in what state it is. (The Proteus website claims it's over 30 years old).

What sort of comparison would you make between programs such as Proteus and Altium versus KiCad?
They're quite different beasts. At least I think so. I'm simply in no situation to afford a EUR 5000 Proteus license  or Altium, which price I don't even dare look up. And when comparing prices, don't start with crippled 200 pin software or 80 sqare cm 2 layer boards. KiCad is no slouch. Some people are asking for updates on the 32 copper layer limit in KiCad. They were using 26 copper layers and were starting to worry about this limit. Others were complaining that KiCad was getting a bit slow on their i9 with a 10.000 footprint PCB. (Or was it hundredthousand?) The story is somewhere in the bug tracker on github. Maybe also migrated to Gitlab when KiCad switched.

[delfinom]

KiCad ain't even 25 year old! It started as a hobby project from a university professor, and it's only really taking off in the last handful of years. Before that it was pretty buggy and development was much slower then now. KiCad is picking up development speed each year.

Yep, we are really picking up speed. The GitLab migration has done absolute wonders. FUCK LAUNCHPAD. What a shitty service that honestly hinders and stunts growth for any project dumb enough to use it. It continues to exist and do untold damage to open source projects.

Same in KiCad. In Pcbnew (that is the PCB part of KiCad ) it already works this way for quite some time. Eeschema (The schematic part of KiCad ) is likely to follow with the next mayor update.

Yep, this behavior is implemented in eeschema in the next major update. You can try it in a nightly.

There was a legacy/history of how eeschema and pcbnew were implemented back in the day. 5.1 was actually an emergency update to convert eeschema partially to the newer rendering framework because the old one got horribly broken on macOS I believe. The next major cleans up a bunch of the remaining inconsistencies.

[PlainName]

It started as a hobby project from a university professor

Well that's a good way to get perverse design decisions embedded in it at the start

But most things probably start like that. Sometimes some higher-up decides that restarting afresh, but this time implementing all the things learned from jury-rigging the original, is a Good Idea. I think the majority of those restarts end up worse, or dead, or just taking forever to get anywhere. Such is life...

What sort of comparison would you make between programs such as Proteus and Altium versus KiCad?

I'm not quite sure what you're asking, there. I suspect any answer would be quite complicated since the position of a product in the marketplace will have an effect. It may be rubbish, but if it's the #1 product then you need to use it if you're going to be part of things. Even if you restrict to purely personal jobs that will never be seen by anyone else, if your PCB house, for instance, understands quite a few native formats but not the product you use, that could be a bit of a kink in the road for you. The likelihood of getting support from somewhere like EEVBlog can be important too

[b_force]

I had my daily dose of comedy again in this topic.

It's just such a thing, that most open source developers don't seem to understand one very important rule.
Even if you software is free and open source, you are still competing against commercial products.
Or in other words, you have to understand the market you're in and how your customers think, need and like.

Most developers (incl KiCad) don't even get so far, but if they do, they always seem step in the same huge mistake.
"We asked our community and therefor..... (fill in)"

Let me tell you one basic thing about marketing, the biggest mistake you can make, is ask your current userbase if they're happy.
Especially for a product/company that still needs to prove himself and want to gain popularity, you should be focused in almost everything except your current userbase.
Be VERY involved why other people are not using your program yet and try to understand why.
Take every bit of critique and feedback dead serious.

Otherwise, especially in the commercial world, you won't even last at all.

The last word I heard almost directly from the KiCad team, was that keeping it multi platform/OS stable is their main focus. 
I still have an headache from facepalming my head.

I lost my faith in KiCad, they just don't seem to be open minded to any feedback and (no offence), their usebase (fans) seem to be even worse.

Someone called Diptrace earlier in this topic, although it still has some major flaws, they actually do take feedback pretty serious and I have seen them making a lot more progress with a team that is a lot smaller.
The fact that it's free and open source is just a extremely poor excuse, there are plenty of examples of free and open source programs that make tons of more progress in even just a month.

[delfinom]

I lost my faith in KiCad, they just don't seem to be open minded to any feedback and (no offence)

We are. We prefer it in the issue tracker with _details_. The feedback that "THIS IS BAD", but never any explanation, or suggestion of a solution or reference doesn't help. The more concrete descriptions, the better.
The wishlist is constantly growing and also being implemented these days https://bit.ly/3091xtv

I personally have watched Dave's eevblog videos with kicad and used that to knock off issues just by watching someone actually trying to use it which is something other videos such as tutorials don't really show you gripes.

And the mention of DipTrace has shown me yes, there actually is another control scheme that is possible, seriously, I have never seen it before and I bet other devs haven't either. Most people will have a legacy of one CAD tool lineage or another and thats it.

But this is what I mean by a description or reference.

The last word I heard almost directly from the KiCad team, was that keeping it multi platform/OS stable is their main focus. 
I still have an headache from facepalming my head.

When both CERN and The Linux Foundation are both major funders of the project....what choice is there? CERN is going full removal of Microsoft internally.
And honestly, the cross platformness isn't too problematic. It really is macOS that struggles but it's usually given it's own hacks for macOS instead of breaking things for Linux and Windows.

Honestly, my biggest headache is all the pre Windows 10 users I would love to kick off. But those Windows 7 diehards just resist progress.

[b_force]

I lost my faith in KiCad, they just don't seem to be open minded to any feedback and (no offence)

We are. We prefer it in the issue tracker with _details_.

I forgot to add some other point, now I know which one.
Sorry, rule two with marketing, don't expect that your customers will exactly tell you what to do in exactly the format you like.

Otherwise I could also just fix my plumbing issue myself, instead of calling for a plumber to fix it, right?
That's the same idea.

You, as a "company" or team have to subtract the main issues from people who are complaining.
Just thinking that they are just there to bash you, is just plain naive and maybe even a little arrogant.
Unfortunately, a lot of coders seem to be only able to think in ticking of buck tracking lists, aka; computer says no. 
Anything else in an other format and they seem totally lost and helpless.

I am very sorry, but if you call KiCad's interface en user experience on par, you VERY clearly have never ever worked with professional EDA's
Maybe that sounds to blunt for some people, but I rather look the reality in the eye.
I am also more than happy to be proven wrong.

[delfinom]

I am very sorry, but if you call KiCad's interface en user experience on par, you VERY clearly have never ever worked with professional EDA's
Maybe that sounds to blunt for some people, but I rather look the reality in the eye.
I am also more than happy to be proven wrong.

I have never said it was on par with anything in terms of quality.

But hey, at the end of the day you can't satisfy all "customers". It is why multiple commerical EDAs exist. Each one meets different needs and in different ways.

[b_force]

I am very sorry, but if you call KiCad's interface en user experience on par, you VERY clearly have never ever worked with professional EDA's
Maybe that sounds to blunt for some people, but I rather look the reality in the eye.
I am also more than happy to be proven wrong.

I have never said it was on par with anything in terms of quality.

But hey, at the end of the day you can't satisfy all "customers". It is why multiple commerical EDAs exist.

Obviously that wasn't directly meant to you, but just a general statement.

[b_force]

In most cad software...

Oh boy! That's a messy can of worms to tip over

It sure is, but for some things you have to find a compromise somewhere.

I like the way Altium implements drag selection: drag down and to the right selects anything completely enclosed, while dragging up and to the left selects anything with a part enclosed.

Same in KiCad. In Pcbnew (that is the PCB part of KiCad ) it already works this way for quite some time. Eeschema (The schematic part of KiCad ) is likely to follow with the next mayor update.
Small difference: KiCad only distinguishes with dragging left to right, or right to left. It don't care about up and down, that is up to you.

Proteus is a good example, having its roots in DOS and dragging that user interface into Windows with it, but over time (25 years or so) they changed to a more Windows-centric interface. Still not quite correct, but a lot better than it used to be.

KiCad ain't even 25 year old! It started as a hobby project from a university professor, and it's only really taking off in the last handful of years. Before that it was pretty buggy and development was much slower then now. KiCad is picking up development speed each year. Give KiCad a few more years (it won't need 25) and then see in what state it is. (The Proteus website claims it's over 30 years old).

What sort of comparison would you make between programs such as Proteus and Altium versus KiCad?
They're quite different beasts. At least I think so. I'm simply in no situation to afford a EUR 5000 Proteus license  or Altium, which price I don't even dare look up. And when comparing prices, don't start with crippled 200 pin software or 80 sqare cm 2 layer boards. KiCad is no slouch. Some people are asking for updates on the 32 copper layer limit in KiCad. They were using 26 copper layers and were starting to worry about this limit. Others were complaining that KiCad was getting a bit slow on their i9 with a 10.000 footprint PCB. (Or was it hundredthousand?) The story is somewhere in the bug tracker on github. Maybe also migrated to Gitlab when KiCad switched.

Sorry, but isn't this obvious?

The whole issue is just all about priorities.
Get the basics done, make it that most people can use it, than work your *ss of to get a proper user interface that even my grandma understands.
Instead of wasting time in auto routers as well as schematic simulators and some multi platform/OS thing that is absolutely not important to anyone except an handful.
the first one is totally useless, the second one you're competing against LTSPice which has an ENORMOUS userbase, or Microcap which is also free nowadays and has been around for many years, the last one can be temporarily fixed with just WINE.

[Cerebus]

... the last one can be temporarily fixed with just WINE.

Sorry, while you're right (IMHO) about much, you're way off base there. That 'temporary' fix would be an almost perfect way of accumulating technical debt. If your aim is to be multiplatform (and that's a perfectly legitimate aim, one that everybody but the poor benighted souls who are stuck with a choice of Windows, Windows or Windows are grateful for) then targetting one platform and using a band-aid to let it run on others is a sure fire recipe for disaster. A lot of things will run under WINE, but won't run very smoothly and tend to have lots of UI glitches and more. You end up patching your code to make it work better with the band-aid, piling ugliness on ugliness, creating tech debt like it's going out of fashion.

I've done my fair share of ports (of commercial software) from Windows to other platforms and it's not pretty, not pretty at all. Designing for multiplatform from the get-go really is the only way to do it without storing up trouble for yourself in the future.

[james_s]

I tried every EDA I could get my hands on about 10 years ago. Decided they all suck, all of them have UIs that are terrible in one way or another. At this point though I know how to use KiCAD so I'd probably just be annoyed if they changed it. Also as someone else said, it isn't Windows software, it just happens to support Windows among other operating systems. I would not expect it to behave exactly like Windows, if it did then it would be inconsistent on other platforms.

[pointhi]

The last word I heard almost directly from the KiCad team, was that keeping it multi platform/OS stable is their main focus. 
I still have an headache from facepalming my head.

Who in his right mind is against cross-platform software? It's more work, but its the only sensible development model to be independent from commercial software vendors without sacrificing most of the userbase.

[PlainName]

I would not expect it to behave exactly like Windows, if it did then it would be inconsistent on other platforms.

Doesn't have to be exactly like Windows, and it's going to be consistent on one or more platforms - that's a fact of multi-platform ability and the platforms be far apart. The point I was raising was that Windows had it about right, so having a great user interface just happens to be like Windows used to be.

By contrast, the way things currently seem to work is to make it not consistent to any platform. Does Linux or Mac have the drag-select-move thing I've been banging on about in here?

And while it is great and useful to be multi-platform, the overriding platform right now is and has been Windows. Linux on the desktop may have its day, but it's not yet (though Microsoft seem determined to push us there). Mac is still an also-ran. So if you have a choice and can't decide which way to fall, surely the way to go is where your largest potential user-base is?

[Doctorandus_P]

Will you please stop banging on the drag-select-move thing.
You've been told, and confirmed by others that it already works this way for quite some time in Pcbnew, and is already implemented in the nightlies for Eeschema.

What do you mean with:

, so having a great user interface just happens to be like Windows used to be.

Windows ain't so great for you anymore?

So if you have a choice and can't decide which way to fall, surely the way to go is where your largest potential user-base is?

I am very happy that I'm able to not use windows.

[PlainName]

You've been told, and confirmed by others that it already works this way for quite some time in Pcbnew, and is already implemented in the nightlies for Eeschema.

1. It may not be in PCBNew but, as mentioned before, we are talking Kicad.

2. I accept it is fixed in the nightlies.

3. I use it because it's an example. Doesn't matter if it was fixed 20 years ago, if one is looking for something to use as an example that we can all relate to, that's what an example is! Give me another example if you're too sensitive for this one.

Windows ain't so great for you anymore?

Are you kidding? Undiscoverable, hidden buttons, invisible window edges... The philosophy is just as bad: a universal interface for phones, tablets and desktops.

[b_force]

The last word I heard almost directly from the KiCad team, was that keeping it multi platform/OS stable is their main focus. 
I still have an headache from facepalming my head.

Who in his right mind is against cross-platform software? It's more work, but its the only sensible development model to be independent from commercial software vendors without sacrificing most of the userbase.

Who said I am against it?

I am just thinking from a practical stand point of view.
If you look at the market, the amount of EE engineers that use anything else than Windows professionally is probably less than 5% maybe 10% at most.
One thing is certain, that market is NOT gonna change any time soon, even if you had the best possible software ever.
So making something cross-platform is very noble an probably more a (50-60s style) ideology, it's just not practical.
And it won't be practical anytime soon.

First gain enough momentum, as soon as you have enough of it, you can change things.
That's exactly the same strategy most successful companies use.
I repeat it once again, even if your software is free, you're still competing against paid software.