Hi,
I’m about to start a new project and I have 8.0 rc3 installed at the moment to try to help out with testing.
Assuming the stable 8.0 release is released before I seriously get into the project am I still better off using the the stable 7.0.11 once it is released ?
I guess my question is - how stable are the new releases ? are they very buggy at the start ?
Current 8.0 RC does seem pretty stable, no major complaints. If I was to start a project now, I’d go for 8.0RC. I’m also pretty sure that if some reproducible, major bug shows up during 8.0 testing, it will be fixed promptly.
I guess my question is - how stable are the new releases ?
Exactly 93,456% stable. Or 93,789%? Damn, I lost my manual “how to measure the stability of software releases”.
some unsorted thoughts:
to get a comparison you could look back to the tranistion v6–>v7 one year ago. Look for kicad v7, all issues with priority high or critical could be relevant. (be warned, the gitlab search function is not very good)
there were some serious issues in the first releases, I think until mid 2023
but: most of these issues affected only edge cases or special functions
so the question “are they very buggy at the start” is not really helpful:
a definition of “very buggy” is missing
it doesn’t help you if it’s generally very stable, but the only remaining bug directly affects your special usecase/workflow
My recommendation:
if stability is your biggest concern: wait 6 month
if you have special workflows/requirements: wait1…2 month, but play already with testprojects or with small projects
if you use only the “most used” functions and are willing to take a very small risk: use v8 now
as proof that I trust my own recommendations: I already use the v8 (rc1/rc2) version since 2 month for all my professional projects.
With previous versions, it always took a few months and bug fix releases for KiCad to become stable. Although I do expect KiCad V8.0.0 to be more stable on release because of the release candidates used now, there is now way to be sure until a month (or two) has passed.
For the plugins, with previous mayor version changes, at least some of the plugins had to be updated to work properly. I am not sure if KiCad has changed in such a way that the same applies for KiCad V8.
The most important fact is probably your own character. KiCad V7 is for the conservatives, while V8 is for the adventurous. Also: A project can always be updated to a higher mayor KiCad version, but once saved in a new version, it can not be opened anymore in an older KiCad version.
Which is why I create a projects directory for each KiCad release and copy from older to newer and then open and update the new version, while leaving the older release version untouched.
For me it is also important whether the plugins and other related tools support the new version… For example I heavily use KiBot so until it works with V8, I will stay with V7…
V8 produces industry standard Gerber, drilling and other fabrication outputs that any PCB house can use. Using native KiCad project files is an extra and theoretically might cause errors due to some version mismatch
Thanks, I was aware, but so far have been just uploading the KiCad files rather than exporting to gerbers and drill files.
I (somewhat cheekily) asked OSH Park on Mastodon what the timescale was for v8 support, and Laen from OSH replied that it could be “as soon as this weekend”
Has there ever been serious consideration for giving KiCAD the ability to save files in a previous version? (Similar to the Office suites, which still give you the option to “Save this document in WORD 97 - 2003 format.”?) I imagine this would improve collaboration among, say, hobby user groups - or a prior client who is still using KiCAD 6.0.
The latest file format can have features that just cannot be described in the older file type.
Even your Word 97 example does not work reliably and I am not a user of the exotic features of Office.
Platform and version problems messing document formatting are a notorious feature.
In a ECAD suite, subtle errors can spoil the project with expensive consequences if you get a batch of broken boards built
Perfect example is in Word if you select “save as old format”, you get a warning that things will break.
And things do break, sometimes horribly with formatting.
In Kicad while we are throwing the idea around of supporting it, things can similarly break and cost you thousands of dollars in damages
For example, say we added teardrop support in v8, and now you want to save as v7. We have no choice but to remove all the teardrops when saving back as the v7 format. The downside is, teardrops are subtle and they may be impacting the manufacturing reliability of the PCB.
We can’t start trying to find ways to emulate teardrops in v7 because that’ll consume a ton of development time to cover every edge case, and actually makes things worse when the v7 saved file is now opened in v8 and it has strange not-teardrop objects acting as teardrops. Lol
In some of the more regulated work I do, if I open a old design in newer software, the entire design will have to get revalidated costing $$$$$$$. I don’t want that. I’ll use the old software just for that old design. Lol.
If I’m going to start spending significant amounts of time in the design in old software, then the question does shift to “well I can just open this in newer software and requalify the design because I’m doing significant amounts of work anyway”.
You keep a copy of Kicad 6 for use with that client, just as you keep a copy of Eagle or whatever else for use with other clients: the cost of having, and keeping, clients.
As for hobby user groups: if you participate in the group, use the same software & version the group uses.
The only real reason I can think of for opening something in an older version is to correct a problem caused by ignorance because the user has not RTFM before saving.
Maybe the yellow warning shown before saving should read:
"This file was created by an older version of Kicad. It will be converted to the new format when saved and will no longer open in the older format "
Maybe the warning should also be changed to cover the whole screen, and bite the user on the nose.