[jitsi-dev] Quality vs Features


#1

The following is a "feature" Request that I submitted to the tracking
system, but was deleted because I did not post it to this mailing list
first. (my apologies)

Jitsi has a lot of great features, and that is great. I want to say up
front that I'm not knocking the features. Unfortunately, jitsi is
really quirky also, and often misbehaves.

I just wanted to submit something for the developers to consider for
developing future versions. Not simply my personal opinion, but a poll
of currently 3000 anonymous people. The question is
"All else being equal, which of the following is more important to a
product's success?"

The poll can be found here
http://gopollgo.com/all-else-being-equal-which-of-the-following-is-more-important-to-a-products-success

And at this time, the votes are:
votes 24% Twice the features of your competitor
votes 76% Twice the quality/polish of your competitor

Perhaps this information can help, and it would be my hope that the
prominence of this can find it's way into the future roadmap of jitsi.
Thank you


#2

Hey Sepero,

I've been meaning to comment on this for a while but couldn't get around
to it. Sorry for the delay.

(inline)

The following is a "feature" Request that I submitted to the tracking
system, but was deleted because I did not post it to this mailing list
first. (my apologies)

That's alright. It happens rather often.

Jitsi has a lot of great features, and that is great. I want to say up
front that I'm not knocking the features. Unfortunately, jitsi is
really quirky also, and often misbehaves.

It would help if you could be more specific about which quirks and
misbehaviours you have in mind.

I just wanted to submit something for the developers to consider for
developing future versions. Not simply my personal opinion, but a poll
of currently 3000 anonymous people. The question is
"All else being equal, which of the following is more important to a
product's success?"

The poll can be found here
http://gopollgo.com/all-else-being-equal-which-of-the-following-is-more-important-to-a-products-success

And at this time, the votes are:
votes 24% Twice the features of your competitor
votes 76% Twice the quality/polish of your competitor

Perhaps this information can help, and it would be my hope that the
prominence of this can find it's way into the future roadmap of jitsi.

Well, it is kind of interesting, so thanks for sharing. Unfortunately it
isn't really that helpful.

While I certainly agree that that bugs are an annoyance to all of us,
and we should keep doing our best to resolve them, the question above is
way too general.

I am sure that when confronted with the above question most users would
think something like "I would definitely prefer a product to work as I
need it to rather than to have new stuff added to it".

The problem is that users stop agreeing as soon as they are asked what
it is exactly that they consider core features that need to be solid and
which ones can be left out.

A lacking feature that matters to someone is pretty much a bug to them.
In Jitsi for example one cannot have contacts belong to multiple groups
even if the underlying protocol supports this. Many people would
consider this to be a bug. Others wouldn't care about it that much and
would consider the "new feature" that allows a password reset to be a
thousand times more important since there server (e.g. jit.si) doesn't
give them another way to do this.

Your previous mail about the single window interface would be another
great example of something that many people may find important and many
others would object to or not care about.

As Jitsi developers we get to see a number of complaints about a number
of things. We also receive a number of requests from BlueJimp customers.
As a result we end up with our own vision of what's important and what
isn't and we try to address issues based on that vision. We really are
doing our best.

Of course, there's no reason at all why our vision should be yours or
anyone else's and it probably won't be :). Luckily Jitsi is open source
and anyone who doesn't agree is free to submit patches for the bugs they
find important.

Hope this helps,
Emil

···

On 14.08.12, 15:45, Sepero wrote:


#3

Hi Emil,
After all this time had passed, I assumed my email was gone with the
wind already. Unfortunately it is difficult for me to file proper bug
reports because I stay busy (as I'm sure you do), and additionally
it's VERY VERY difficult to get debug info from average computer users
(aka my friends). My only desktop system is a computer running Ubuntu
Linux. I would prefer to use Jitsi as my instant messenger, but when
it comes to sound quality, it is just cannot match a proprietary
messenger which I will not name. The strange thing is that Jitsi
delivers poorer sound quality and takes more bandwidth.

I have a friend "A" who uses MSWindows. When I call her, and the audio
often sounds robotic. When I screen share with her, the screen sharing
often crashes after a few minutes.

I have another friend "B" who also uses MSWindows. I talked him into
installing Jitsi. When I called him, the call connected, I could hear
him, but he could not hear me. We tried multiple calls. He uninstalled
soon after.

These are bugs and quirks make me very reluctant to ask anyone to
install Jitsi. My friends will not be happy for me to recommend a
program which does not work decently at it's fundamental purpose.

-If a program is missing a feature that a person needs, they may
decide not to use the program, but they will likely come back and try
again in the future if they had an overall positive experience.
-If a program has all the features that a person needs, but it crashes
or fails to work in fundamental ways that give them a negative
experience, they are unlikely to come back.

A lacking feature that matters to someone is pretty much a bug to them.

Others wouldn't care about it that much and
would consider the "new feature" that allows a password reset to be a
thousand times more important since there server (e.g. jit.si) doesn't
give them another way to do this.

I know exactly what you're saying, and as a fellow developer I
understand you face a lot of challenges. To identify "quality", it may
help to prioritize items in terms of a hierarchy. To go from "general"
to more specific, this is how I might possibly construct such a
quality hierarchy, from most important to least important:
-text/audio/video connection and transmission quality
-polishing UI
-documentation (in program and website)
-program/account configuration
-adding UI components

Now I know that the Jitsi team operates on hierarchy somewhat similar
to this already, because some bugs are rated as more important than
others. I have not seen any such hierarchy listed on the website, so I
don't know if it exists, or if the developers are just using a system
based on intuition.

I also understand that some things must be judged on a case by case
basis. (such as your example of the jit.si password change). A
hierarchy like this serves best as a guide, more than a hard-fast
rule.

I can tell that you are doing the best you are able to, because you
would have never responded if you didn't care, and users like myself
really appreciate that. I am sincerely grateful for the whole jitsi
team and I've been trying to switch to your program for years, ever
since it was called "sip communicator" :slight_smile:

Hopefully one day I can make the full switch, because I much prefer
using open source (and not being spied on), and Jitsi is without a
doubt the top oss competitor. Primarily because of it's
multi-operating system capability.

In my humble opinion- 1) fix the fundamental connection issues 2)
polish the UI 3) get it on Android. And I think people will be
migrating to Jitsi faster than we can even imagine.

All the best,
Sepero

PS.
Is there a way to turn on "recording" all program operation into a
file? If so, then my programming illiterate friends could send me that
file, and I could use it to file a bug report.

···

On Fri, Aug 17, 2012 at 11:00 AM, Emil Ivov <emcho@jitsi.org> wrote:

Hey Sepero,

I've been meaning to comment on this for a while but couldn't get around
to it. Sorry for the delay.

(inline)

On 14.08.12, 15:45, Sepero wrote:

The following is a "feature" Request that I submitted to the tracking
system, but was deleted because I did not post it to this mailing list
first. (my apologies)

That's alright. It happens rather often.

Jitsi has a lot of great features, and that is great. I want to say up
front that I'm not knocking the features. Unfortunately, jitsi is
really quirky also, and often misbehaves.

It would help if you could be more specific about which quirks and
misbehaviours you have in mind.

I just wanted to submit something for the developers to consider for
developing future versions. Not simply my personal opinion, but a poll
of currently 3000 anonymous people. The question is
"All else being equal, which of the following is more important to a
product's success?"

The poll can be found here
http://gopollgo.com/all-else-being-equal-which-of-the-following-is-more-important-to-a-products-success

And at this time, the votes are:
votes 24% Twice the features of your competitor
votes 76% Twice the quality/polish of your competitor

Perhaps this information can help, and it would be my hope that the
prominence of this can find it's way into the future roadmap of jitsi.

Well, it is kind of interesting, so thanks for sharing. Unfortunately it
isn't really that helpful.

While I certainly agree that that bugs are an annoyance to all of us,
and we should keep doing our best to resolve them, the question above is
way too general.

I am sure that when confronted with the above question most users would
think something like "I would definitely prefer a product to work as I
need it to rather than to have new stuff added to it".

The problem is that users stop agreeing as soon as they are asked what
it is exactly that they consider core features that need to be solid and
which ones can be left out.

A lacking feature that matters to someone is pretty much a bug to them.
In Jitsi for example one cannot have contacts belong to multiple groups
even if the underlying protocol supports this. Many people would
consider this to be a bug. Others wouldn't care about it that much and
would consider the "new feature" that allows a password reset to be a
thousand times more important since there server (e.g. jit.si) doesn't
give them another way to do this.

Your previous mail about the single window interface would be another
great example of something that many people may find important and many
others would object to or not care about.

As Jitsi developers we get to see a number of complaints about a number
of things. We also receive a number of requests from BlueJimp customers.
As a result we end up with our own vision of what's important and what
isn't and we try to address issues based on that vision. We really are
doing our best.

Of course, there's no reason at all why our vision should be yours or
anyone else's and it probably won't be :). Luckily Jitsi is open source
and anyone who doesn't agree is free to submit patches for the bugs they
find important.

Hope this helps,
Emil


#4

Hey Sepero,

Hi Emil,
After all this time had passed, I assumed my email was gone with the
wind already. Unfortunately it is difficult for me to file proper bug
reports because I stay busy (as I'm sure you do)

Unfortunately it is even more difficult for us to fix them without
proper bug reports. Of course having a bug reported does not guarantee
it will be fixed but if it is reproducible and affects of lot of people
then chances are pretty good.

and additionally
it's VERY VERY difficult to get debug info from average computer users
(aka my friends).

Our log files are pretty easy to obtain:

https://jitsi.org/faq/logs

My only desktop system is a computer running Ubuntu
Linux. I would prefer to use Jitsi as my instant messenger, but when
it comes to sound quality, it is just cannot match a proprietary
messenger which I will not name.

We use SILK, the same codec as Skype, so audio quality should be comparable.

The strange thing is that Jitsi
delivers poorer sound quality and takes more bandwidth.

Interesting. Could you please tell us a bit more about your test
environment. I generally get between 10 and 30 KBps with SILK for a
great quality. Are you using another codec?

I have a friend "A" who uses MSWindows. When I call her, and the audio
often sounds robotic.

Available resources may affect quality. You may ask her to disable echo
cancellation and noise reduction. This could help a bit.

When I screen share with her, the screen sharing
often crashes after a few minutes.

We'd need to see logs from both sides next time this happens. It would
also help to know exactly what you mean by "sharing crashes". Does the
application complete crash and exit? Does the session end? Does she stop
seeing your desktop? Does this only affect the key and mouse events?

I have another friend "B" who also uses MSWindows. I talked him into
installing Jitsi. When I called him, the call connected, I could hear
him, but he could not hear me. We tried multiple calls. He uninstalled
soon after.

Would need to see logs.

These are bugs and quirks make me very reluctant to ask anyone to
install Jitsi.

Well, again, reporting them and helping us track them down is a
necessary pre-condition to having them fixed. It is also a lot more
helpful than conducting and submitting a general survey regarding users'
preferences between stress and comfort. We really need no convincing as
far as fixing the above is concerned. We need help.

My friends will not be happy for me to recommend a
program which does not work decently at it's fundamental purpose.

-If a program is missing a feature that a person needs, they may
decide not to use the program, but they will likely come back and try
again in the future if they had an overall positive experience.
-If a program has all the features that a person needs, but it crashes
or fails to work in fundamental ways that give them a negative
experience, they are unlikely to come back.

I don't necessarily agree with the above, but that's not really relevant
here anyway. Again, you do not need to convince us that bugs need fixing.

A lacking feature that matters to someone is pretty much a bug to them.

Others wouldn't care about it that much and
would consider the "new feature" that allows a password reset to be a
thousand times more important since there server (e.g. jit.si) doesn't
give them another way to do this.

I know exactly what you're saying, and as a fellow developer I
understand you face a lot of challenges. To identify "quality", it may
help to prioritize items in terms of a hierarchy. To go from "general"
to more specific, this is how I might possibly construct such a
quality hierarchy, from most important to least important:
-text/audio/video connection and transmission quality
-polishing UI
-documentation (in program and website)
-program/account configuration
-adding UI components

Thanks for sharing. We'll keep that in mind.

Now I know that the Jitsi team operates on hierarchy somewhat similar
to this already, because some bugs are rated as more important than
others. I have not seen any such hierarchy listed on the website, so I
don't know if it exists, or if the developers are just using a system
based on intuition.

Right. We don't use such a hierarchy. We treat every issue on a case by
case basis.

I also understand that some things must be judged on a case by case
basis. (such as your example of the jit.si password change). A
hierarchy like this serves best as a guide, more than a hard-fast
rule.

I can tell that you are doing the best you are able to, because you
would have never responded if you didn't care, and users like myself
really appreciate that. I am sincerely grateful for the whole jitsi
team and I've been trying to switch to your program for years, ever
since it was called "sip communicator" :slight_smile:

Thank you very much for your encouragement, support and kind words! We
do very much appreciate this! We are really sorry for the problems that
people may be experiencing with Jitsi, and we are doing our best to
address them as we can. We are also very grateful when people help us
with the fixes.

Hopefully one day I can make the full switch, because I much prefer
using open source (and not being spied on), and Jitsi is without a
doubt the top oss competitor. Primarily because of it's
multi-operating system capability.

In my humble opinion- 1) fix the fundamental connection issues 2)
polish the UI 3) get it on Android. And I think people will be
migrating to Jitsi faster than we can even imagine.

Again, thanks for sharing. This is not far from our own vision.
Hopefully we'll soon be in a state that completely satisfies your needs,
so stick around!

Cheers,
Emil

···

On 17.08.12, 19:19, Sepero wrote:

All the best,
Sepero

PS.
Is there a way to turn on "recording" all program operation into a
file? If so, then my programming illiterate friends could send me that
file, and I could use it to file a bug report.

On Fri, Aug 17, 2012 at 11:00 AM, Emil Ivov <emcho@jitsi.org> wrote:

Hey Sepero,

I've been meaning to comment on this for a while but couldn't get around
to it. Sorry for the delay.

(inline)

On 14.08.12, 15:45, Sepero wrote:

The following is a "feature" Request that I submitted to the tracking
system, but was deleted because I did not post it to this mailing list
first. (my apologies)

That's alright. It happens rather often.

Jitsi has a lot of great features, and that is great. I want to say up
front that I'm not knocking the features. Unfortunately, jitsi is
really quirky also, and often misbehaves.

It would help if you could be more specific about which quirks and
misbehaviours you have in mind.

I just wanted to submit something for the developers to consider for
developing future versions. Not simply my personal opinion, but a poll
of currently 3000 anonymous people. The question is
"All else being equal, which of the following is more important to a
product's success?"

The poll can be found here
http://gopollgo.com/all-else-being-equal-which-of-the-following-is-more-important-to-a-products-success

And at this time, the votes are:
votes 24% Twice the features of your competitor
votes 76% Twice the quality/polish of your competitor

Perhaps this information can help, and it would be my hope that the
prominence of this can find it's way into the future roadmap of jitsi.

Well, it is kind of interesting, so thanks for sharing. Unfortunately it
isn't really that helpful.

While I certainly agree that that bugs are an annoyance to all of us,
and we should keep doing our best to resolve them, the question above is
way too general.

I am sure that when confronted with the above question most users would
think something like "I would definitely prefer a product to work as I
need it to rather than to have new stuff added to it".

The problem is that users stop agreeing as soon as they are asked what
it is exactly that they consider core features that need to be solid and
which ones can be left out.

A lacking feature that matters to someone is pretty much a bug to them.
In Jitsi for example one cannot have contacts belong to multiple groups
even if the underlying protocol supports this. Many people would
consider this to be a bug. Others wouldn't care about it that much and
would consider the "new feature" that allows a password reset to be a
thousand times more important since there server (e.g. jit.si) doesn't
give them another way to do this.

Your previous mail about the single window interface would be another
great example of something that many people may find important and many
others would object to or not care about.

As Jitsi developers we get to see a number of complaints about a number
of things. We also receive a number of requests from BlueJimp customers.
As a result we end up with our own vision of what's important and what
isn't and we try to address issues based on that vision. We really are
doing our best.

Of course, there's no reason at all why our vision should be yours or
anyone else's and it probably won't be :). Luckily Jitsi is open source
and anyone who doesn't agree is free to submit patches for the bugs they
find important.

Hope this helps,
Emil

--
Emil Ivov, Ph.D. 67000 Strasbourg,
Project Lead France
Jitsi
emcho@jitsi.org PHONE: +33.1.77.62.43.30
http://jitsi.org FAX: +33.1.77.62.47.31