[DaveBlake]

"Lies, dam lies and statistics...."

The truth is we really don't know how many users we have, what version they use, what features they use and for how many hours a day. The scraper data is going to include all the testing and trials of new installations, and miss all the older installs that aren't updating their library, and people like me that don't bother to scrape. The Yatse data is a view of what Yaste users do, maybe we can scale it, maybe we can't because there could be an inherent device bias. Persoanlly I think the Yatse data is as near we can get to an understanding of version use in the real world, and with reasonable caution, we should pay heed to it.

We delude ourselves if we think all users are rushing to upgrade. Some will, some are despirate for a big fix or new feature, others will just stick with what they have if it works for them. That is fine as long as we are sure there are not other reasons that discourage users from upgrading.

But sometimes we might want to learn to appreciate what work has been done and what is working/has already been achieved. Even if it's not the perfect solution.

I have huge respect for all the work that has gone into Kodi, both for the code and the people. Gratitude is a good thing in all areas of life. I must say that I have recieved plenty of appreciation and politeness from other team members and users. But is it also clear that all is not well within the team, and I am not sure if appreciation is enough to fix it.

Some people feel very abused and frustrated, some people do seem to respond in an egotistical and malicious way (and just saying they mean no malice is not enough). I think someone else commented that all sides have reason to be the way they they are, and I do not want to get drawn into a sides thing. Can we try and learn from it?

[Koying]

And an important thing about the stat I gave since it was not clear

There's 1 line for each combination of precise Kodi Version / OS build.

Well, if anything, that proves that numbers can be made to say anything

I'd be interested in:
- your global ratio android/others
- you ratio 15+ android/15+ others
- your ratio <15 android / <15 others

Thanks

[Tolriq]

Reading dave post I need to maybe add something

I do love Kodi and respect all the work that is made and will be made and I do respect 99% of the Kodi Team (Yes I should avoid this part but after so many years I can't change )
My only concern is about external coders handling (and in a certain way the global vision that may exist or not but is clearly hidden to users)

About upgrade reasons or not, let's talk about non numbers but users feedback, When Helix was out there's was a big API change for PVR that allows better EPG grabbing. As most users where complaining about speed in Yatse for that part I said hey upgrade it will be better, and for the majority the answer was : I won't update to a .0 version it will be bugged for sure.

And indeed for 14, we faced the infamous windows crash

About stats since there's a page change please be sure to read previous page

And a little reminder about the stat report I used before : http://forum.kodi.tv/showthread.php?tid=...pid1796456

I could also say something like Yatse have not becomed what it is by luck and it's very high rating is also not luck, stats and numbers and user listening is the key to success, and yes Kodi is successful but it could be so much more I dream of a world with Kodi everywhere

[Koying]

That's a bit of an unfair generalization

Granted, hence the smiley

I might have misremembered - wouldn't be the first time - but IIRC builds I installed via .apk downloads from here updated via the play store. Am I imagining that?

Yeah. That's a decision from Google to do automatic updates of all apps present on GPlay, even if they were sideloaded. It actually makes some sense, but it's unfortunate that users have zero granularity in controlling this (it's auto-update or not).
A workaround would be to use a different certificate to sign the sideloaded apk and the GPlay one. I think I remember I proposed this, but don't remember the outcome.

[trogggy]

I read the explanation at the time (it's all coming back to me) - I understand why it happened, just the piracy reference seemed a bit off. And I think I must have a blind spot for smileys.

Anyway, I agree about the certificates. If that meant I could keep a play store version but also try nightly builds without too much of a faff it would be even better. I've suggested it in the past - for now the nearest I can get is to use SPMC (which is also brill, thankyou) alongside kodi.

[Tolriq]

I'd be interested in:
- your global ratio android/others
- you ratio 15+ android/15+ others
- your ratio <15 android / <15 others

Thanks

Quickly done with split on / (Cause some problem on some linux release but should not affect too much)
Os is checked with Kodi infobools : System.Platform.XXXX

For sunday 3rd ratio are : (I usually do stats on sliding 30 days but I already had exported this)

Global android ratio : 11% (11% of Yatse users use Android)
15+ Android ratio : 14% (14% of Yatse users with Kodi 15+ are on Android)
14- Android ratio : 6.7% (14- means 14 included) (6.7% of Yatse users with Kodi 14+ are on Android)

Another interesting stat is : Android 14- ratio vs all android : 24%

Meaning "if we can extrapolate but I know some do not want :p" that there's a lot more non Google Play release used. Surely those are the bad pirated box or SPMC

[Tolriq]

Yeah. That's a decision from Google to do automatic updates of all apps present on GPlay, even if they were sideloaded. It actually makes some sense, but it's unfortunate that users have zero granularity in controlling this (it's auto-update or not).
A workaround would be to use a different certificate to sign the sideloaded apk and the GPlay one. I think I remember I proposed this, but don't remember the outcome.

Different certificate can have side effects for many things.

What most devs do when dealing with this problem (Mainly for the Amazon store problem, GooglePlay update install from Amazon) is to play with version code numbers.

Those are what triggers the update, for Amazon I add 990000 to the version code so Google Play can't update the Amazon version. (The opposite is still possible but it's user choice).

This could work for this case, manual APK build should have a way higher version code (even with same version name) and problem solved.

[Koying]

Global android ratio : 11% (11% of Yatse users use Android)

Yep, my point exactly. Last Kodi stats I saw would show android users > 50%. So Yatse population != Kodi population.

BTW, if our stats are anywhere like correct, 70% of our users on latest version is very good.
Windows 7 still amount for 50% of desktop market share.
Android 5.0+ is only 25% of Android market share.

[Koying]

This could work for this case, manual APK build should have a way higher version code (even with same version name) and problem solved.

Yep, probably a better solution, indeed.

[Razze]

"Lies, dam lies and statistics...."

The truth is we really don't know how many users we have, what version they use, what features they use and for how many hours a day. The scraper data is going to include all the testing and trials of new installations, and miss all the older installs that aren't updating their library, and people like me that don't bother to scrape. The Yatse data is a view of what Yaste users do, maybe we can scale it, maybe we can't because there could be an inherent device bias. Persoanlly I think the Yatse data is as near we can get to an understanding of version use in the real world, and with reasonable caution, we should pay heed to it.

We delude ourselves if we think all users are rushing to upgrade. Some will, some are despirate for a big fix or new feature, others will just stick with what they have if it works for them. That is fine as long as we are sure there are not other reasons that discourage users from upgrading.

But sometimes we might want to learn to appreciate what work has been done and what is working/has already been achieved. Even if it's not the perfect solution.

I have huge respect for all the work that has gone into Kodi, both for the code and the people. Gratitude is a good thing in all areas of life. I must say that I have recieved plenty of appreciation and politeness from other team members and users. But is it also clear that all is not well within the team, and I am not sure if appreciation is enough to fix it.

Some people feel very abused and frustrated, some people do seem to respond in an egotistical and malicious way (and just saying they mean no malice is not enough). I think someone else commented that all sides have reason to be the way they they are, and I do not want to get drawn into a sides thing. Can we try and learn from it?

Even if you not bother to scrape, it still will update the addon, as it's installed by default unless you disabled it updating?
The rest seems quiet spot on. The appreciation thing was more about code and doing PRs, not really aimed at the current team status. It's not healthy to always concentrate on the bad things, that's what I'm trying to say.
Hard to motivate people if all you do is say everything is bad and everything they do is bad too


The real/only question regarding the non upgrades, is why? Speaking feature/bug wise, other things are probably out of our hands. (Maybe your favorite skin stopped support... )

[Tolriq]

Global android ratio : 11% (11% of Yatse users use Android)

Yep, my point exactly. Last Kodi stats I saw would show android users > 50%. So Yatse population != Kodi population.

BTW, if our stats are anywhere like correct, 70% of our users on latest version is very good.
Windows 7 still amount for 50% of desktop market share.
Android 5.0+ is only 25% of Android market share.

Well I'd like to see how Android version could be more than 50%, Kodi is great on Android, but I really doubt about this number, pirated box are not that high level.

About stats from Razee as I tried to explained the 30% are far from real as not the correct indicator

[Koying]

Well I'd like to see how Android version could be more than 50%, Kodi is great on Android, but I really doubt about this number, pirated box are not that high level.

Well, there is already a factual 3M out of a realistic 10M. Taking into account the non-gplay boxes, it seems plausible, at least...

[trogggy]

The real/only question regarding the non upgrades, is why? Speaking feature/bug wise, other things are probably out of our hands. (Maybe your favorite skin stopped support... )

If all you're doing is watching local media then there's probably little reason for many people to upgrade if everything's still working. I generally upgrade asap, try nightlies, run lots of versions. I'm constantly messing / tweaking. But for my actual usage I could be several versions back and I doubt anyone but me in the house would notice.
I'd guess a lot of older builds are also ones that have been set up by someone else. If you didn't install it in the first place then there's an element of 'If it ain't broke...'
And then there are the pre-installed boxes.

Android is a different case, as upgrades are pretty much forced on users if you use the official / play-store version.

[Tolriq]

Well, there is already a factual 3M out of a realistic 10M. Taking into account the non-gplay boxes, it seems plausible, at least...

We could debate a lot but active install from GPlay is not active users, far from it, and if numbers from tmdb addon then again biased information for global active users.

Anyway as Dave said we do not know, I do give information do what you want with them, when I report user feedback, it's ignored, when I report numbers they are said to be false or irrelevant.

I guess burying head in the sand can work until a truck pass at your exact position

I'll now return to my old position where I try to ignore what happens under my nose.

[Koying]

We could debate a lot but active install from GPlay is not active users, far from it, and if numbers from tmdb addon then again biased information for global active users.

Well, that lower both numbers, but keep the ratio more-or-less the same, I guess.
So, being very conservative, that'd still be 30% of Kodi android users globally, vs. 11% of Android Yatse users.

It's still far too much a discrepancy to take conclusions from Yatse stats alone, which was the original point.

But yeah, at the end of the day, we are clueless because we receive zero stats from Kodi installs themselves.
So taking action regarding a supposed low-upgrade rate based on half-random stats (because there are so many assumptions) wouldn't probably be wise