Interesting reading found on Dave’s forum, commenting on Altium’s 2021H1 report to their investors.
For anyone interested, it’s EDA/Altium section, " The clown is here to stay".
First time after long-time, steady and double-digit growth, Altium’s sales dropped by 4%.
Among comments, there are many voices against Altium’s move to cloud based business model.
There are also voices that confirm KiCad’s great potential in taking the major role in the hardware industry.
Quote from jmw: “I wonder if they’ve internalized the precipice they are approaching. Altium will go the way of Borland. Nobody (outside of miserable non-competitive niche industries) is paying for compilers today, and in 5-10 years nobody is going to be spending thousands for an ordinary PCB EDA package. KiCAD is on the path to becoming the GCC of the hardware world.”
I think it’s a good moment to say big Thank you to our KiCad developers, for their involvement in making the greatest open EDA package.
Don’t read too much into sales drop. 2020 was a year of contraction and besides a few industries that boomed (signage for instance) others had to initiate furlough schemes and cut investment.
If growth has been replaced by contraction, a company won’t invest in an additional Altium license or upgrade their present license.
As Altium has shifted to a cloud based annual subscription model, a drop in sales means more than that. Stop paying and you lose access to existing designs.
Sadly, this could be a prediction of how many electronics companies shutter in the remainder of this year after furlough stops
Funnily enough, embedded electronics is precisely one of those niches. KiCad has a long way to go to compete with Altium for the high end (where Altium itself is not really a top contender).
Which is most likely why a lot of software companies switch to this model. After all it is even harder to change to a competitor if your current provider has your data.
But it also provides benefits to the users. After all if you do not own the data then you do not need the IT infrastructure and support in house (especially a benefit for medium-sized companies).
What is however clear is that certain industries can not use cloud based tools. But this can be circumvented by offering on premise solutions (not sure if altium does it already or plans to do it).
I would again not read too much into this as well. My guess is that a lot of companies are just a bit more careful with which employees have a licence. Questions like “you only use it to look at stuff so why do you have a licence?” will most likely more often be asked in a time of economic crisis (at least this is what happened at my company with our eagle licences – which is why i no longer have one).
And certain non-critical departments will potentially be asked to switch to something cheaper (like eagle or potentially also kicad).
And just to make it clear: yes there will be a lot of companies that will go under in the next years as a effect of the current crisis. But the fact that altium lost sales is not on its own evidence of this.
I’m voting on the current pandemic causing a economic crisis, there is absolutely many companies still on shaky grounds. I know of a few small time defense contractors alone that are near the point of closing up shop soon and they are supposed to be recession immune businesses hah
I doubt KiCad has taken over many Altium seats. What it has done is displaced the low end products that used to be in the shareware space.or advertised in the electronics magazines in the old days.
We’re doing lots of embedded programming, and use the toolchain based on GCC.
Proprietary compilers are becoming less important, as embedded programming is heading towards standard architectures.
At firs KICAD will erode the market for low-end software, probably over time will also take over part of the mid-end market. Highend will probably remain for highend tools and not necessarily Altium.
We’ll see what will come in the future.
It’s not so simple to compare something with GCC. It works as a simple analogy “KiCad may become as popular in its own field as GCC is in its own field” but not in any other more complex way. KiCad can’t be compared with GCC (or Clang). The reason is simple: C++ and C are highly standardized languages. GCC doesn’t have a file format to compete with other file formats.
Another reason is that GCC isn’t an end user application. It doesn’t have a GUI per se. End users very rarely use GCC directly, and then it happens with simple command line. On the other hand there’s no one standard GUI or frontend for programming, people use VS, Eclipse, vim/emacs etc.
The closest comparison would be with office applications. They are also GUI applications, complex pieces of software, collection of tools to edit different documents and they have their own file formats (or at least some native file formats like ODF). In that area Open/LibreOffice haven’t been able to replace MS Office.
Where the comparison with office software doesn’t work is that EDA is used by highly specialized people, engineers, and it will always be a niche market when compared to office tools. Even as a hobby it has a quite high barrier to start with. Most people who ever use desktop/laptop computers will at some point use office applications and O/LOffice is good for that. It came pre-installed with this computer. If KiCad would become pre-installed for everyone, people would become angry.
I don’t think that’s true. I contract, writing embedded software, and in the last seven years I’ve only come across one company that uses anything other than gcc. Most people are using the Eclipse
based freebie tools these days.
My prediction is that there will be a lot of interest in EAGLE to KiCad and Altium to KiCad importers in the upcoming years, when people notice that they cannot open their old projects anymore and that they also cannot acquire a permanent license, so their only option is to pay monthly for something they rarely use … or jump ship.
I’ve seen this effect before. Autodesk turning 3ds max subscription-only was one of the best things that ever happened to Blender, because suddenly companies started using it and started dedicating resources to Blender training, external services, and paid support.
Good news, this makes KiCad’s position even stronger, IMO.
It’s possible that companies will use KiCad for creating simpler designs, leaving expensive Alitum licenses only for advanced projects; while being able to open these KiCad made designs in Altium any time they need to.
Some time ago I dropped the idea of writing both importers and exporters for KiCad. The Idea is that if someone wants to convert from X to Y, and it can be done with KiCad, then after a while they will scratch their head and ask themselves why they would either use X or Y.
Writing an importer for KiCad is a lot of work, but most of the work is in reverse-engineering the other format, (or at least getting familiar with it), decoding that format with software, and getting familiar enough with KiCad’s internals to to map entities to the KiCad format. I’m guessing that once you have obtained that knowledge, then writing the software to convert the other way is a relatively simple job.
If you can translate both ways, you have a better platform for highly automated testing, and you can also make use of test files made previously and imported in KiCad.
You can only automate testing after manually testing it first, unless you can figure out a way to programatically open your file in the target system. I haven’t found a way in cadstar at least. In these cases automatic testing just helps avoid regressions but not development.
Even so, the biggest effort is in generating test designs to find all the edge cases. For example in cadstar you can only have zones in one layer but in KiCad 5.99 you can have it in multiple layers. Your zone test designs made for the importer will be incomplete if used to test the exporter as you won’t be exercising what happens with the incompatible features.
