This all presumes that people are reading the blog with every release, and upgrading hass every two weeks. For a developer this might be normal practice, but most people would strongly prefer to ignore their home automation system as much as possible. As a result, they only tend to upgrade when something requires it (see: Harmony update for a recent example). Then, you’re left wading through dozens (hundreds?) of breaking changes strewn across several blog posts. This is not at all user friendly.
3 Likes
Well, OpenHAB only updates twice a year, but that is a long time to wait for official fixes. It is based on Java and halfway migrated to Eclipse Smart Home.
I honestly do not get how this text would have helped me with my initial post.
The first paragraph tells us there will be warning when i used keys which are not allowed (Which i don’t)
The secend paragraph is about lovelace custom cards and a filename change which does not affect me (Oh by the way, here is an perfect example for the attitude, eat or die. They only tell us, it has changed, after it had changed… And already broken peoples installations. And there is no reasoning on why this had to happen)
The third and last paragraph starts with “for custom component developers” which i clearly am not. So i decided it is not ment for me. That they hide the reasoning for almost all changes in this blog post, was in no way clear from the text around it.
We are mainly complaining about breaking changes in core components and just got to custom components most recently as DavidFW1960 brought up a pending change for the future.
Thats the reason i still stick to hass.
I know python much much better than java and java error messages are just a pain in the butt.
FYI, that “future change” about custom components was released yesterday.
my post was in reference to your much quoted complaint…
which was a reply to this link
In which I posted that it was on the main blog post. In it’s own section as a major breaking change.
I guess I could say that this comment may have been out of place.
But I stand by my other statement as it was linked in the main blog.
Sorry, did not see it. Then it is a recent change 
By the way, this time i think the release notes are much better and much clearer. If they were like this every time i guess i would not complain 
2 Likes
Thanks for your complain to allow us have a better release note 
1 Like
This is a self inflicted wound!
3 Likes
Really? I didn’t make these breaking changes.
I am ready for a slating, but I don’t see the issue here.
HA is not yet on version 1. The breaking changes are as a consequence of trying to get to v1…ie out of beta. Like it or not, HA is still considered beta. The fact it is so stable leads many to forget this.
Before I update I always look at the breaking changes. Or I did. For a while I have been running a dev instance and a prod instance - not practical for all, but it works for me. I now install the betas and then iron out all the issues I will have in my prod. My update today took 5 mins with no issues, because I had the issues last week on a system that doesn’t affect my life around the house.
I understand not everyone is able to understand the breaking changes immediately, but how about waiting until v 0.xx.1 or o.xx.2 before you upgrade and checking these forums for issues? Or the HA github? All you need to do is check the changes, check the github for any custom components, and you should be ok.
This isn’t Microsoft or Apple - the releases are tested, but it is free software in beta. Remember that and just keep an eye on the breaking changes for all releases.
I’ve yet to have a change that has rendered my HA inoperable for more than an hour or two…and this is beta software.
Enjoy!
2 Likes
for all those who say:
HA is not version 1 yet, and we are beta users, etc.
if you still keep changing and adding stuff after 5 years (with over 1000 people working on it), and dont have a clear goal to what must be adchieved before first release, i dont see that it will ever be a stable product.
at some point you need to say: at that date i want that adchieved and then we call it a release.
but at this point its still: lets see where we go, what we can add, how we can improve, and maybe we will come to a first release at some point.
for those who are hoping for a first release and then things can be stable: that can take another 5 years, or maybe 10, who knows? not even a dev can tell you when first release will be reached.
in my eyes thats because first release and stabelizing things is not a main goal.
4 Likes
I agree and disagree with you.
Home automation is in its infancy in general. Yes you have the big names out there in Nest, Amazon etc but these are cloud based. They are also a compromise in general and don’t talk to each other.
HA is local and trying to integrate all these different systems into one ‘hub’.
On the other hand it is so stable now that we can use it as a home automation hub. I agree that you need to draw a line at some point and say this is now production, but with 1200 odd 3rd party components this isn’t easy. Imagine life without HA…breaks or not.
I think that over the next few months the breaking changes will reduce as the major ones are all for the future stability of HA.
We all contribute to the system by reporting the breaks and the fact that this forum is so active with strangers willing to help others speaks volumes.
there will always be 3th parties breaking things.
in my eyes there should be a split between core HA and components.
core HA should be stable and soon released and components should be possible to update when needed.
because the whole thing is in 1 single package, updates are needed way more frequent.
2 Likes
Yes. You knowingly chose to use beta software and chose to skip several releases because you were too lazy to keep reasonably current.
If you want fewer updates, try OpenHAB. You will find that releasing every 6 months is a good idea until you need to live with brokenness for that length of time.
1 Like
The trouble with that assertion is that you are suggesting Home Assistant should be a high-maintenance solution. Everyone getting into Home Assistant should expect from day one that they are going to have to be making constant, never ending changes to their configuration just to keep the lights on.
4 Likes
This whole thread is because they are reorganizing to modularize things more.
That is why, for instance, SmartThings needs to be reconfigured in 0.89. When you realize the long term goal, it makes more sense.
the point is that the long term goal keeps changing before first release.
i still remember the time that the devs were at the point where they said (or actually it was mostly balloob): NO, there wont be translations!
and so i can sum up a whole lot of other stuff on what the opinions and goals have changed in the last 2 years.
the long term goal from 2 years ago, is completely different from the 1 there is now, and probably it will differ from the goal there is in 2 years.
there is no real release goal, there is a goal to which direction HA should go.
but at some point people are going to be impatient. at some point you need to give people something thats stable and release.
and in my eyes that point is actually way overdue.
i stopped following every update about 2 years ago, but that also has his downsides, because when you then NEED to update, you have a very hard time finding out everything that did break.
but its really unwanted to shut down my house every 2 weeks because i need to stay up to date.
1 Like
Applied the “if it ain’t broke don’t fix it” and stayed on 0.73. Before I knew it HA went to 0.82 and in the end it was quicker installing 0.82 and starting from scratch as the upgrade from 0.73 to 0.82 was a disaster for me.
Learnt quite a bit about HA in the process and the biggest lesson was don’t fall too far behind.
Unlike some users who stuff their install with every conceivable component I gradually add to HA as I need it. This also helps managing the breaking releases phenomenon and make it a little easier.
Maybe the model of a stable version for the everyday user v the beta for testers and bleeding edge junkies might be something worth considering some time in the future.
I dont think saying there is no direction is very postive
1 Like