... not working ...

users
#1

... Thanks :slight_smile:
              maybe now more will realize what we're saying :wink:

BRAVO Anne-Ology!!

#2

        ... Thanks :slight_smile:
               maybe now more will realize what we're saying :wink:

This is simply unrealistic. For anyone who has any experience with programming this would be known. No offense but with a ratio of 100,000:1 or more users to developers, the idea that we would just squash all bugs and stop releasing new versions isn't realistic at all and thus why developers wouldn't respond to this recommendation. If you want to help I suggest taking some C++ classes and getting involved with the code. Most of us are volunteers who do this with our spare time, I hope you all keep that in mind

Regards,
Joel

#3

        ... Thanks :slight_smile:
               maybe now more will realize what we're saying :wink:

This is simply unrealistic. For anyone who has any experience with programming this would be known. No offense but with a ratio of 100,000:1 or more users to developers, the idea that we would just squash all bugs and stop releasing new versions isn't realistic at all and thus why developers wouldn't respond to this recommendation. If you want to help I suggest taking some C++ classes and getting involved with the code. Most of us are volunteers who do this with our spare time, I hope you all keep that in mind

Regards,
Joel

Another problem for all programs in wide release is wide hardware variability in the Windows and Linux worlds especially when compared to Macs. There could be a very odd hardware/driver interaction that was never discovered in alpha, beta, or release candidate previews with specific hardware combinations.

#4

Let alone regressions which are just accepted as a part of software development. Ultimately, mistakes happen, and when a software code base has been transferred left and right and had a ton of people randomly working on it (which is exactly what happens with open source software), a lot of the job of a developer coming onto the project is just playing "catch up" and guessing and what a previous developer was attempting to do. Again, I highly suggest taking a few C++ classes and then it'll become apparent that the idea that we should stop everything and get every single bug squashed (>5,000), is not a realistic stance. We should and we are (I guarantee this) doing everything in our power to prioritize bugs and take care of those bugs that are

a) most annoying

b) affecting the most users

c) resulting in data loss

We are a young project and this is a goal that has been set. Being young, this is a goal, not a fact. If you're interested in seeing how much work is done on a daily basis, just follow gerrit (our code tracker), or sit in IRC and look at the incredibly brilliant conversations that happen to find solutions to many of the problems that are being reported.

Just to give another point, we are averaging more than 5 new reports PER DAY. Our QA team is a group of volunteers no more than 7 or 8 strong. Each of these bugs has to go through a long process just to verify, ensure that it's not a duplicate, communicate with the user who reported it, and then priortize it. That's JUST getting the bug confirmed, then it gets put into the stack where a very small group of dedicated developers tackle them, one by one. A single bug can take a week + to tackle (that's 40+ hours). Let's say the average bug takes 10 hours (a massive understatement), that's 50,000 hours worth of work to tackle the 5,000 or so confirmed bugs.

Seeing these off hand remarks about how we should develop the product is disheartening. I wish that more people would take a class at their local community college, or take a free online course, and start to put their thoughts to work on our code.

Regards,
Joel

#5

Hi Joel,

I think nobody blames the dev team for not working enough; in contrast all appreciate the work done by the dev team, being it new features or bug fixes.

For my understanding the concerns and opinions expressed here are more related to the ratio of devs being allocate to new features versus bug fixes. I think most of us also understand that first the number of devs is limited, and second the decision on how many devs on new features and how many on bug fixes is not an easy one. It is a management decisions, which can turn out right or wrong. And for management decisions one needs a lot of information. Speaking for myself I would regard these comments as additional information.

In this sense, please regard the messages in this forum as additional information, which could.should help decision makers to make decisions. And should a decision turn out to be wrong, there is no problem in changing course. The worst thing a manager can do is, not to make a decision. Without decision there is no change to the situation and thus no chance for improvement.

Let me thank here again all devs doing great jobs in developing and improving LO.

ROSt

Maybe I take C++ classes when I retire.... LO would be worth to do.

#6

Hi :slight_smile:
+1
I think that it's precisely because we see so much great development being done that we get excited about it and want more.  Also the fact that people can post bug-reports and even non-coders opinions do get listened makes us feel like we can "off-load".  So, a lot of the criticisms are not necessarily about LO specifically but about IT in general.

ROSt talks about management decisions and hints about those being imposed on a workforce because that's the easier way for most of us on the Users List to understand it.  I suspect that most of us either do now or have at some point worked in traditional offices in mainstream management structures.  In LO i imagine the equivalent is leadership rather than management.  Inspiration and initiative rather than "just following orders".  Personal investment and interest in achieving goals.  It is leading to some impressive results.

Regards from
Tom :slight_smile:

#7

I don't mean that a single manager makes all decisions and the team has to follow. The word "manager" could and should be understood also as "management team", "decision making team", "leadership", etc.

I am not familiar with the structure of the dev team and decision making processes. Thus, I used simply "manager". At the end it is the dev team bringing LO forward. Let me repeat myself, but I respect these people a lot and I am very grateful for all the work they did and do.

#8

Hi :slight_smile:
Yes, but rather than learn new words for all those sorts of things i thought it was smart to re-use words that we ARE familiar with in order to give a close approximation to what is going on.

LO is a meritocracy which suggests there are leaders leading.  There is some sort of management structure that has some sort of hierarchy but one that can respond sensibly to people's changing lives rather than being stagnant and brittle.

We don't need to know the exact details and just using words that are similar enough is good enough for us on this list.  If people want details then they can probably find them out easily enough by joining the devs list or looking around their wiki and places.  It's not hidden it's just too much for me to think about right now and their are tons of ways they could have structured it sensibly to deal with the type of work they do.

The main point is that devs are roughly equal to rock gods imo while we hang around helping people with their tickets and point them to the bar. 
Regards from
Tom :slight_smile:

#9

This is going to go on my list of QA stats that I prepare for our team. I'm not sure how I'll get the data BUT when I do I suspect that 90+% of our work goes towards bug fixing, 10% or so goes towards enhancements. Once I get the data together I'll make sure to send it out to the user list. One good way to know that most of our work goes towards bug fixing is to sit in LibreOffice dev chat and watch as the bot lists the patches being pushed, I RARELY see anything but bug fixes. Another good indicator is looking at FDO and seeing that most enhancements are left untouched (unfortunate but we just don't have enough developers).

Regards,
Joel

#10

Another important point is that a lot of our "new features" are about compatibility. It is expected by our users and our contributors that our product be compatible with the competitors. This is unlike our competitors who routinely want to ignore what we are doing to make the office suite better. So when our competitor makes a change, we want to ensure that if you try to open the document(s) in LibO, for the "most part", it works.

It is definitely a hard decision and I hope that more people start tinkering with the code, I'm not a programmer by profession but a bit of work and a lot of patience, and some great help from the dedicated developers has resulted in me submitting a couple patches and a few more are on the way :-D. Don't forget that developing isn't the only way to help, we have marketing, UX (look of our product), documentation, QA, etc... that all need people. Even if it's only an hour a week, it helps, I guarantee it.

Regards,
Joel

#11

Hi :slight_smile:
That was interesting to hear about the Bot scrolling through lists of freshly completed bug-fixes.  I think it might be good to hear about the percentages of what is going into the 3.7.0 specifically and then compare that to the 3.6.5 (or whatever the old branch is by the time the 3.7.0 is about due) but i guess that if anyone is really keen they could just look for themselves just before the 3.7.0 is due.  It might be interesting.

I think your input to this thread has changed a few opinions here.  It's been interesting to hear from someone on the devs lists
Thanks and regards from
Tom :slight_smile:

#12

3.5.7 just came out, so that should be the last of the 3.5 line, unless someone decides to make an update or something.

So, 3.7.0 will be worked on soon and is due out the second week in Feb 2013. Yes that is over 3 months away, but it would be nice to know what the real differences between 3.5.7, the 3.6.5 version, and 3.7.0 version.

What is changed between the last of the 3.5 line and the 3.6 line by the time 3.6.5 comes out. What will be changed between that version and the 3.7 line.

I know there are release notes and such, but it would be nice to see a side by side chart with the changes so you can compare the different versions/lines with each other better.

For the conservative users 3.5.7 will be it until 3.6.x's version gets to the "conservative point". But the more "advanced" users will be using 3.6.x till 3.7.x comes out.

That is the life cycle of the two line development cycle.

#13

This is a very interesting information. I am not really surprised by the ration indicated here. I exepcte finding the cause of a bug takes much more time than final fix. I am grateful to the dev because I know myself how difficult it is to find a bug in a complicated SW.

Question: What means:
- bot lists
-FDO

#14

The idea expressed below is a very good one. I too would appreciate such a comparison table. Placed somewhere at the release notes.

#15

In IRC (chat client) we have automatic robots that routinely (ten's of times a day) say what patches have been submitted to gerrit (the service we use to keep our code). Every time a patch is submitted the robot (bot) spits out an automated message saying that a person (with a name) has submitted a patch and explains what that patch does.

FDO = free desktop . org (f d o). Ans is where you can report bugs pertaining to the libreoffice project. If you want to see a list of confirmed bugs look here -- you'll see immediately how overwhelmed a small project can get:

https://bugs.freedesktop.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&list_id=147729&product=LibreOffice&query_format=advanced&order=opendate%2Cchangeddate%2Cbug_status%2Cpriority%2Cassigned_to%2Cbug_id&limit=0

That is 3,956 bugs and climbing (there are an additional 1,200 or so that are reported but not confirmed yet...the QA team confirms the bugs so developers can focus on coding)

Regards,
Joel

#16

Thanks for the information.

Theh bug list is quiet long.... Hope we get 100 new devs eager to fix bugs.... or should I retiere earlier and learn C++....

#17

Hi :slight_smile:
Ahhh, now i understand what QA is.  I think other projects call it different things such as the Triage Team and stuff like that.  I think a few people on this list could almost definitely help and maybe we could help push people in that direction.

A few people have asked about how to become a dev and i usually point them to just the Easy Hacks but QA is a good way of getting hands-on quickly without actually having to instantly learn coding.  You clearly build-up a better understanding of the infrastructure used by LO devs and perhaps it helps you as you are learning coding by giving you examples of where it's most needed.  Right, next time i will point people towards QA.

Many thanks for all the snippets of information about all this!
Thanks and regards from
Tom :slight_smile:

#18

Hi :slight_smile:
Ahhh, now i understand what QA is. I think other projects call it
different things such as the Triage Team and stuff like that. I think a
few people on this list could almost definitely help and maybe we could
help push people in that direction.

+1. We do more than just confirming bugs in QA but it is a big part of what
we do.

A few people have asked about how to become a dev and i usually point them
to just the Easy Hacks but QA is a good way of getting hands-on quickly
without actually having to instantly learn coding. You clearly build-up a
better understanding of the infrastructure used by LO devs and perhaps it
helps you as you are learning coding by giving you examples of where it's
most needed. Right, next time i will point people towards QA.

+1 please do, you can have them even email me directly as I coordinate a
lot of that stuff.

Regards,
Joel

#19

Hi :slight_smile:
Just one thing.  LO is not particularly small and it's definitely not new.  The original code was called Star Office and was developed around a decade or so ago.  After a couple of years Sun took it on and called it OpenOffice.org and then TDF took the code and called it LibreOffice.  So there are probably chunks of the code and ways of doing things that date back to the last century!  Hence why you still see references to "soffice" if you look in your task-manager or systems processes.

Now that a lot of the old irrelevant comments and stuff have been almost cleaned out (allegedly) it might be possible to focus more on streamlining and using more modern approaches in some areas.  It should definitely be easier to find your way around the code and i think that is going to have some big impacts on the effectiveness of any work done by the devs.  The Pita stuff is nearly done.  Time for some fun!  (or have i got it wrong again?)
Regards from
Tom :slight_smile:

#20

I would say the code is "getting better". I've only been on the project a
few months so I can't say how much progress we've made but I know that we
still have a build time of 4+ hours which says a lot. The code is still a
bit tricky, not enough comments left by previous developers, etc...

Also, I would argue that "our project" (meaning LibO, we don't use LO :wink: ),
is quite small still. We have a very small team of developers who routinely
(>1 patch per week) contribute, a slightly bigger team that submits at
least one patch a month, and then a couple hundred that have submitted 1 or
2 patches in the course of several months (this includes myself ). We could
probably double our number of developers doing 1 patch or so a month and
still be in need for quite a few more.

Regards,
Joel