And then imagine how users might complain that the version 7 DNP feature isn’t honoured in version 6 and someone didn’t read the import failure log and ordered the wrong component. I really don’t see this as a high priority. Old versions of KiCad scan be downloaded and a team just needs to agree on what version to use and all move together.
Of course, when there is a maximum difference between versions of 10 years or more, but there is one year of difference
It’s only in theory… In practice, from kicad to altium converts well to the reverse side there are not many errors and a lot of manual editing… Starting with inscriptions and layer names ending with fonts… For an amateur level or simple devices, you can probably crash, but you can forget about the transition of commercial developers… Similarly, the picture looks like with other programs such as libre linux freecad, etc. They can do something but as a complete replacement or so to speak, there is no full-fledged competitor…
The reason why you end up with this is that open source aims to solve a problem at hand. Not do deep research into how you make a system thats fully workable. In this case the real enemy of kicad is its 2d drawing layer, it is built around a line drawing engine.
Now it turns out that the minimum needed feature set of a line drawing and a shape filling system is different. In practice you can probably live with a system which only works on polylines but it makes for a very nasty programming problems. Simply put you need to be able to make items and eventually holes in items. While you can kind of bork this by drawing a connecting bridge to each hole it becomes very hairy very fast. Implement a poly line primitive with arcs and holes and things become much much easier to do and everybody would be using it (better add b or bezier splines).
Rather surprisingly this kind of system also makes back porting easier. Because all you need is a translator that translates things into the underlying system at worst you just dump the geometry. But kicad can not do this because its still not easily usable feature complete on the 2d drawing layer.
Altium does not have this problem.
This is not a criticism, just a feature of incremental design changes. In practice i find that KiCad is probably the second or third best open source software out there. This says more about the others than it says about KiCad though. Its just that I feel KiCad is at the cusp of the network effect so it can suddenly become the best there is.
1 Like
Open source should not mean flawed or not functional. For the end user, convenience speed and functionality are important… At the moment, I cannot do my work quickly and qualitatively with the help of kicad librecad freecad etc… Of course, you can adapt to using part of the plugins to read the converters part of the manual work question for what?) This also applies to simply operating systems replacing win with linux… They proudly shout that win is bad but they themselves use win applications))
Obviously not, and its hardly the goal anywhere. But in reality it means that those who know what they want and are willing to put up resources get what they want. All software is flawed to some degree its not worth crying over.
Open source just has different incentives. Sometimes it works sometimes not so much. If it does not do what you want then do what you do when you encounter same problem elsewhere adapt or discard.
Tell me how you can adapt, for example, linux to work with 2d graphics and replace autocad, for example? Or how can kicad replace altium’s cam gerber editor? I’m not talking about minor flaws in software and these are global things… Do you know what the developers of kicad answer? It is too difficult))) I answer well make a separate application for example gerbv say no it is too difficult))) At the same time, a bunch of plugins repeating the functions of cad itself is not difficult)))
This thread is going down hill fast and is on the verge of being closed. If the software doesn’t meet your needs, pitch in or move on. There are alternatives.
2 Likes
Well i have replaced autocad a few decades ago with something way more appropriate for me. But no you cant expect replacement for X, it does not happen ever. What you can do is build a entirely new workflow that can be better.
Also they didn’t say too difficult, they said that they think their resources are better spent elsewhere. Give enough resources and you can have nearly anything.
And i agree it should be closed @hermit
2 Likes
This is my point of view. Not on kicad but the things in general have a evolution process:
- First step: hangup, fails and lot of bug
- Promise: this can be great
- I can use it. ( personal usage)
- I can work with it ( profesional usage)
- i need it to my work ( if dont i am out of job market)
Internally we can locate kicad where we want, and check what is missing for the next stage. And even ask, we want go to the next stage? What is need for the next stage?
i want to remark that we are speaking about kicad comparing with altium, at top level. And not with ms-dos autotrax or “Smart” ( for more older people hehehe). This let in background a message: Kicad can be the best.
Hermit, i think in a different way. Few months ago i was owner of a small factory with a 15k to 20k products sold. Here i learn few things, 99% simply make their choice as you said, the 1% of people suggest/demand new functions or things. Of course some things are possible and another does not. But as soon i add it that the customer demand the sales grow a 5 to 20%. This mean that the most people do as you said ( use or not under the convenience) but is very fewer the people that say why is not using or purchasing.
I go back to the start: If is possible save in backwards compatibility, will be great for me (and i think that will be great for more people). I read nice contribution about save as primitives, and i like it.
if is not possible, is ok.
2 Likes
I think one of the points is that if someone wants to create a file format converter from KiCad 7 to 6 (or even to 5) it’s realistically possible. Conversion from 7 to 6 wouldn’t be terribly difficult because the format is plain text, even human readable, there are no major changes between the format versions, and it can be scripted with any widely used programming language. Basically it would take only one new interested person who has technical inclination and some programming skills. It wouldn’t even need much maintenance afterwards because the file formats won’t change anymore for these same versions. There would be need only for a next major version converter after a year, and for that the maintainer would need to track down the changes between the major versions but the code architecture would already be there. Only one volunteer who thinks this is worth it!
The other important point is that the current development team won’t do it, period, and there’s no need for it. Only one new volunteer, or one company who can hire a summer employee. No need to use the precious time of the core developers of KiCad.
The problem though is that users of such a converter may have unrealistic expectations of it. For sure there will be attributes in v7 files so some semantics cannot be expressed in v6, after all devs are adding new featues all the time otherwise why augment the format.
So while you probably can create something that opens in v6, you might find important bits missing that would damage the design, and this jeopardises collaboration.
Well, that’s even more true for, say, other EDA to/from KiCad conversion. Typically these are used once and then the design process is continued in the new format and the problems are fixed as the first step. To be honest I would never try a process which does lossy conversion again and again. All KiCad versions are free to use and the older version should be chosen and fixed as the standard if it’s the lowest common denominator. But sometimes it would be worth doing the conversion once to an older software version.
The converter will also have to be rebuilt every year.
Probably within a week of a new release, the questions will start: When will the converter be ready?
Just as the questions have been “When will “this or that” Plugin be ready”?
And if someone does build a converter, but only once or twice, then stops, who is going to calm the mob?
You, of course, because you are writing here. 
2 Likes
I left myself wide open for that comment. 
Yes, a compatibility layer will require some format stability. This is not bad, is good, and is the quality certification process.When you try to adapt to iso9001 and newer this is an objetive. This talk about of maturity of a product.
Sadly i only know about perl
I think all that’s to be said already has.
1 Like