šŸ“ˆ Forum activity check

Community Playground General Discussions
#41

I think you misunderstood me. I wasnā€™t talking about rushing anything. If you read my various posts and this one too. I am saying you should pace to a higher rate of release which would lead to much greater software quality by taking smaller steps each time. Right now we are observing long intervals between releases and very large amount of changes which are poorly tested and as far as the community can seeā€¦ poor quality. You donā€™t need to rush anything to release bad qualityā€¦ Look at the vera history

If you take smaller stepsā€¦ say only a handful of commits between releases and allow your community to beta test it so you can fix them before general public releases which would be at a much higher rate, you would get a greater community participation and firmware quality.

As I see 7.0.30 is way too ambitious to be taken as a single step for example. The content could and should have been split into 3-4 releases. The work @edward has done on the luaUPnP program alone could have been 1-2 general public releases which would have greatly improved stability of a large number of customers. It is all about how you manage itā€¦ You are taking the risk to have much bigger release increments, much more difficult to troubleshoot for your teams and lower quality overall, as well as a less active community.

For comparison, Home Assistant has a public release every 5-10 daysā€¦ The state of their software stability is leaps and bounds better than vera, granted with much greater resources. If you could get to a rhythm of even 1/3 of that, you would be in a much better place.

Hubitat likewise when I was following them had a new firmware release every week. No matter how small the changes were. They enabled us to test it and if buggy, get reverted or fixed the following week.

1 Like
#42

Sorry, I just have to interject on this one. One of the reasons I stopped using HomeAssistant is that it seemed every fourth of fifth release contained breaking changes that sent me scrambling to fix automations, or UI, or something. I routinely found things that didnā€™t work as documented in both core and components. On top of this, component updates are inconsistent: some fast, some slow, some good, some bad. There are currently 822 open issues in Github for Hass. The development team fixes things when they feel like it. Balloob lords over everything. In several cases past, where several users pointed out completely conflicting behavior vs the documentation, it took months to get fixes, and a few issues that I participated in are even today not yet fixed. Patches abound, and often are lost and need to be redone, or are invalidated completely and need to be reconceived, by core updates. And among many of the component developers with whom I had interaction (including pyvera, the library used to connect HA to Vera), things were often regarded as ā€œworking good enoughā€ and there was little that could be done to motivate the improvements necessary to make things right (for reference, the lead developer of pyvera even told me he doesnā€™t really use Hass himself, he just uses his library for his own stuff, so Hass is a secondary customer in his world).

I would not say thatā€™s leaps and bounds better than Vera. At best, it may be on similar footing. But I never even came close to getting my Hass automations as deep and as reliable as what I can do on Vera (eating my own dog foodā€“I use my own plugins, so I know if I can get done what I need, so can this community, or Iā€™ll fix it).

You know that you and I agree completely on a fast release cycle; weā€™ve discussed that many times, so no argument there at all. And yes, adamantly, 7.0.30 has become a perfect example of why they must.

But I would not hold up Hass as a shining example of software quality, and I would not be happy if Veraā€™s release behavior was like that of Hass.

Hass is a community-developed product. It literally has +1,600 contributor who have made contributions at various levels, some continuously, some just once. This type of product naturally will have a lot more discussion. It also runs on a variety of OSs, and a larger variety of hardware, and with several different Z-Wave sticks and other connectivity components. So naturally, thereā€™s going to be a lot more questions. Thereā€™s also going to be a lot more problems. So they have a lot to talk about.

3 Likes
#43

You both have a point.

Butā€¦ asking something on HASS, gives me a real answer with a real solution. And next release it is fixed.

Veras last updat was about a year? And the next one is in beta for ages nowā€¦ and I only know by reading so e posts on this forum and not even those from Vera themselves.

#44

+1 on that.

I left my HA install on a particular version as it was working nicely. I needed something on a newer version which was 5 points ahead of my own and as soon as I updated all hell broke loose on the breaking changes front.

That coupled with some really unhelpful ā€œwise guysā€ on the forum turned me off HA completely.

I came back to VERA after a two year lay off and have to admit the forum is a breath of fresh air in comparison. No smart asses. Just users trying to help each other.

If anyone wants the frantic update cycle HA is definitely the place for them. I think I had enough when I started to see plugins for cat flaps and bin collections.

1 Like
#45

No HA is not ideal. I was trying to illustrate what could be.
I have had no unfixable issues with them and yeahā€¦ I dislike the breaking changes too which is the reason why I donā€™t update unless I need to. The point is that the release rate they are at enables them to fix and implement in much safer increments and they are not shy about reverting code.
Also, I ran my own branch for some time and implemented my own fixes when the main branch was too slow to implement them. I actually contributed some fixes there too. I spend very little time on their forum. Only on an as need basis. It is just too fast to keep up with but if you are looking for a live forumā€¦ this is it. It is an extreme vera does not need to go to (neither do I think that it can since it is not community driven) but could certainly learn some from.

#46

I have to say, Iā€™d be remiss if I did not acknowledge our (my and other developersā€™) contribution to that. We asked for a lot to be fixed. And to their credit, they addressed almost everything we asked for. Now reading all of this and thinking about the process, @rafale77 and I actually broke our own rules that we are promoting by doing that. I think we were just so caught off guard by having one of the Vera Devs listening and responding to every single thing we said, the attention weā€™ve asked for for so long, that when we got it, we asked for too much, and perhaps we should have backed off as well.

So, for the current Beta that I know many of you are waiting on, donā€™t blame only Vera for the timing. We hold some responsibility for that as well, and I will take that as a lesson to go into the future.

But I also want to be sure you (the community) understand, moreso than ever before, the Vera Devs have been involved with us and listening to us. It gives me high confidence that when we find the right rhythm, strike the right balance, we will get that consistently fast progress we are looking for.

6 Likes
#47

Bravo!!! (extra text because forum software demands it.)

2 Likes
#48

Just to illustrate how much progress was made, I am posting a screenshot of my home automation interface which is a combination of vera-openluup and home assistant. Note the uptimeā€¦ I have a very high confidence I can beat my previous record with this 7.0.30 test runā€¦ This hopefully will get some more interest going in the community. :slight_smile:

45
451833Ɨ973 90.3 KB

2 Likes
#49

What flavour of HA is running there.

#50

HA is tedious to configure but once done it is very configurable and powerful. This is a vanilla version in which I configured iframes to have the other home automation UIs on the same menu within HA.

#51

Is it hass.io running on a Docker or one of the other install methods.

#52

It is the full version. I am not aware that hass.io is any different though.

#53

I run HA in a venv. Not the preferred choice of the HA forum but it gives me most control.

#54

I think we will have to agree that the forum has now come alive! :grin::laughing::smiley:

2 Likes
#55

For this topic :stuck_out_tongue_winking_eye:

#56

We see a heartbeat :slight_smile:

#57

Since 7.30 beta release and preview updates on the iOS and Android app the forums have become far more active. I see this as a good sign for Vera.

1 Like
#58

Vera? Sorin? Melih? are you still alive?

I see about 3-6 new topics a day and about 10 repliesā€¦

#59

Youā€™re still doing something odd.

Iā€™ve had well over 40-50 replies today. Granted not that many topics but thatā€™s that.

Nothing from @Sorin for a while now. One post I saw from @melih some time last week IIRC

C

#60

I think I alone already average about 10 replies a dayā€¦ And I donā€™t think I just talk to myself.

Actually I just checkedā€¦ I average over 12 posts a day without counting my post edits.

1 Like