[steering-discuss] Base - a new mailing list?

HI :slight_smile:
Base needs to have some documentation to help devs understand what it's trying to do.

With the other apps it's much easier to have an overview even without having any documentation but Base is a lot more complex.  Also there needs to be some discussion between devs and other people to decide what to use instead of java, and the direction to take generally.  Apparently the default HqSql back-end is fairly troublesome but there is a new project "MariaDb" that is a drop-in replacement for MySql that might be much better as a back-end.  MariaDb is basically almost the entire community that worked on MySql including some of the original people plus a load of new people that joined since the project broke free from Sun/Oracle.  Anyway it needs a small team of people to try out different back-ends to see which are viable as back-ends and which might be easiest to have as the default one.  Perhaps bringing their recommendation back to steering-discuss to make an informed vote.

As i understand it the current infra-structure is leading to Base gently crumbling away through neglect.  Experienced devs steer well away from Base.  New ones need to learn more understanding about coding or about Base or about both.  Every time an individual joins and is keen to work on some aspect of Base they realise it's a completely tangled mess and they would be the only person working on it so they get discouraged and give-up.

Yes it would be great to have input from the entire devs list but
1.  They are not interested in Base
2.  The non-coders that want to do non-coding work on Base (eg Documentation, perhaps marketing, perhaps design of aspects of the UI that only appear in Base, the wizards and so on) would be unable to understand.

Regards from
Tom :slight_smile:

HI :slight_smile:
Base needs to have some documentation to help devs understand what it's trying to do.

With the other apps it's much easier to have an overview even without having any documentation but Base is a lot more complex.  Also there needs to be some discussion between devs and other people to decide what to use instead of java, and the direction to take generally.  Apparently the default HqSql back-end is fairly troublesome but there is a new project "MariaDb" that is a drop-in replacement for MySql that might be much better as a back-end.  MariaDb is basically almost the entire community that worked on MySql including some of the original people plus a load of new people that joined since the project broke free from Sun/Oracle.

MariaDB is like MySql, so it is a nice and powerful database, but too
heavy weight(*) for a 'default' back-end. SQLite sound like a much
better match for the task.

(*) by too heavy weight, I'm referring to installation, configuration,
need for started-task/daemon, need for periodic maintenance etc...

Anyway it needs a small team of people to try out different back-ends to see which are viable as back-ends and which might be easiest to have as the default one.  Perhaps bringing their recommendation back to steering-discuss to make an informed vote.

That's not how it works... voting is not the answer... doing is. I'm
pretty sure that if somebody step up to make a viable support for a
self-contained back-end, it will fidn its way into the code by
consensus, without the need for a 'vote', or getting the steering
commity or future BoD involved in any ways.
BTW this is already listed as an 'Easy Hack' :
http://wiki.documentfoundation.org/Development/Easy_Hacks#default_to_SQLite_not_HSQLDB_in_Base
So clearly this is not a technical/political disagreement, but just a
matter of someone stepping up to do the work...

As i understand it the current infra-structure is leading to Base gently crumbling away through neglect.

It is not an infrastructure problem. it is a lack of interest problem.

Experienced devs steer well away from Base.  New ones need to learn more understanding about coding or about Base or about both.  Every time an individual joins and is keen to work on some aspect of Base they realise it's a completely tangled mess and they would be the only person working on it so they get discouraged and give-up.

That is true for every part of the product. the effort needed to find
your way in the code is high, and base is not special in that regard.
I seriously doubt that it is easier to find your way around writer
than it is to find your way around base.

Yes it would be great to have input from the entire devs list but
1.  They are not interested in Base

That is indeed a problem, but I fail to see how starting to section
the community vertically ( vs horizontally like the ml are now) will
help

2.  The non-coders that want to do non-coding work on Base (eg Documentation, perhaps marketing, perhaps design of aspects of the UI that only appear in Base, the wizards and so on) would be unable to understand.

I don't understand what you mean with that paragraph...

Norbert

Hi :slight_smile:
It's not to restructure the entire TDF existing infrastructure!  What is there works extremely well except in the case of Base.  We clearly don't need separate lists for the other apps as they are all doing fine as they are both in documentation and in keeping devs that start work on them.

It is good to hear that the default back-end issue appears to have been decided already although i thought there were problems with Sqlite too.  Certainly MariaDb sounds too heavy for a default.  Anyway i didn't mean to start discussion here as most of the people that have expressed interest in Base in the last few months are not on this list.

Base loses people almost faster than it attracts them but it does attract an odd person or 2 occasionally. 
Regards from
Tom :slight_smile:

Hi,

As i understand it the current infra-structure is leading to Base gently crumbling away through neglect. Experienced devs steer well away from Base. New ones need to learn more understanding about coding or about Base or about both. Every time an individual joins and is keen to work on some aspect of Base they realise it's a completely tangled mess and they would be the only person working on it so they get discouraged and give-up.

I doubt you will create more interest with yet another mailing list. What I would propose: Actually start *working* on Base, may it be documentation, code or any other aspect, and *then*, when there is enough discussion traffic, we can talk about another list. All other program modules currently have no need for their own list, so I would not make an exception for Base, but rather see if there is really a demand...

Florian

Hi :slight_smile:
This has not worked in the last year and is unlikely to work in the near future.  I have seen people start work on Base before but because any lists that they posted questions in were more interested in almost anything other than Base they got discouraged or distracted. 
Regards from
Tom :slight_smile:

Hi,

This has not worked in the last year and is unlikely to work in the near future. I have seen people start work on Base before but because any lists that they posted questions in were more interested in almost anything other than Base they got discouraged or distracted.

about how many people are we talking? How many people would like to work specifically on Base, and in which areas?

Florian

Hi :slight_smile:
Do you mean how many expressed an interest and tried to give it a go during the past year or do you mean how many at any one time or do you mean how many right now?
Regards from
Tom :slight_smile:

Hi,

Do you mean how many expressed an interest and tried to give it a go during the past year or do you mean how many at any one time or do you mean how many right now?

the rationale behind my question is that I want to avoid a mailing list where only ten people or so contribute actively and regularly. I personally just don't see a need for this list at the moment, and would rather prefer if people discussed it on the other lists; only if that does not work then, we can think about a new list. First the work, then the list, I'd say. :slight_smile:

Florian

Tom Davies wrote:

Do you mean how many expressed an interest and tried to give it a
go during the past year or do you mean how many at any one time or
do you mean how many right now?

All of that would be useful to know.

A list does not magically make a community appear. I'd suggest
either discuss@ or libreoffice@lists.freedesktop.org to attempt to
coagulate interested parties - libreoffice@fdo has the advantage
that Cc-ing unsubscribed parties permits them to answer to the list
w/o moderation. Feel free to Cc (or Bcc) the folks you'd think might
be interested, by proposing some specific things to do (like the
developer docs you mentioned).

Once there's people working on stuff, and still want a separate
list, feel encouraged to come back with a proposal here.

Cheers,

-- Thorsten

Hi :slight_smile:
Done that.  Been there.  It didn't work.  Base is dying. Can we just admit that and remove it from LO?
Regards from
Tom :slight_smile:

I think part of the problem is the rise of client server databases with the
internet. It's a bit of an irony because to start with OOo used the
principle of connecting to a database rather than including the old Addabas
that was with StarOffice. Snag now is that even if the use of Base is
minority it's difficult to withdraw it without upsetting them.

Ian
www.theingots.org

Hi :slight_smile:
Done that. Been there. It didn't work. Base is dying. Can we just admit
that and remove it from LO?
Regards from
Tom :slight_smile:

I think part of the problem is the rise of client server databases with the
internet. It's a bit of an irony because to start with OOo used the
principle of connecting to a database rather than including the old Addabas
that was with StarOffice. Snag now is that even if the use of Base is
minority it's difficult to withdraw it without upsetting them.

Ian
Sent from my Android Smartphone.
www.theingots.org

Hi :slight_smile:
Done that. Been there. It didn't work. Base is dying. Can we just admit
that and remove it from LO?
Regards from
Tom :slight_smile:

Why not find a way to integrate connectivity to all the major databases such as mysql and MsSQL servers?

I think part of the problem is the rise of client server databases with
the
internet. It's a bit of an irony because to start with OOo used the
principle of connecting to a database rather than including the old
Addabas
that was with StarOffice. Snag now is that even if the use of Base is
minority it's difficult to withdraw it without upsetting them.

Ian
Sent from my Android Smartphone.
www.theingots.org

Hi :slight_smile:
Done that. Been there. It didn't work. Base is dying. Can we just admit
that and remove it from LO?
Regards from
Tom :slight_smile:

Why not find a way to integrate connectivity to all the major databases
such as mysql and MsSQL servers?

This was the method in the early days. I assume it still works. At the
weekend at the Apache Bar Camp I was talking to PostgreSQL developers who
are very interested in better integration with OOo/Libo. We are likely to be
working together on training and certification so there are possibilities to
get some funding to this development but its going to take a little time.
They are applying for FP7 funding through the EU and we can complement that
with Lifelong Learning projects.

Tom Davies wrote:

Done that.  Been there.  It didn't work.  Base is dying. Can we
just admit that and remove it from LO?

Hi Tom,

why do you think it's dying? And removing it from LO is not an
option for me. People using it need to step up, and start getting
involved - and I'm sure they will. Free software is a lot about
scratching your itches - if something does not work properly, get
your hands dirty & try fixing it. We'll all happily answer code
questions over at libreoffice@lists.freedesktop.org

Cheers,

-- Thorsten

Tom, Ian, please move the discussion over to the discuss@ list. If
there's any code questions (and there were people interested in the
long-dormant postgres-connector), that should go to
libreoffice@lists.freedesktop.org.

Thanks,

-- Thorsten

Hi :slight_smile:
Base does support a lot of different back-ends but it needs to have a default one.  The current default seems to be quite troublesome so people are often told to use something else such PostGreSql or MariaDb / MySql.

MariaDb is the same as MySql except that it is developing fast and has almost all the MySql community including the original developers.  The community were even more unhappy with being under Oracle than they were under Sun so they forked off and formed a new organisation so that they could push through a large number of bug-fixes and developments that Sun / Oracle had been blocking for years.  I did mention this earlier in this thread but i know it's difficult to keep track of issues like that when this list is not focused on Base so i guess i have to mention it again.  MySql is in roughly the same state as OOo before Apache got involved.  MariaDb is a drop-in replacement with a much stronger future.

Clearly people on this list don't know much about usign Base.  When you open Base the first thing it asks is which back-end you would like to use and there is a drop-down that includes the various back-ends mentioned and more but has a default of HqSql.  Again i guess it's something that people on this list might not be aware of so it's a opint that will probably have to be mentioned several times in the course of any discussion about Base in this or any of the other existing lists.

If the PostGreSql people could be encouraged to send some devs to work on Base then that would be a huge help and would greatly help tighter integration with that particular back-end.  It would be nice to include other people that are interested in working on various aspects of Base, eg doumentation, devs, maybe design etc but not all those people are on this list.

Perhaps one way would be to cross-post any discussion about Base so that all the lists got any post about Base?  That would neatly avoid having to set-up a new list and still reach the various different people :slight_smile:

Regards from
Tom :slight_smile:

Hi :slight_smile:
I am sorry but there is NO suitable list to discuss Base.  Almost no-one that has expressed any interest in working on Base is on the list you discuss and they are not willing to join a high-traffic list that knows nothing about using Base.
Regards from
Tom :slight_smile:

Hi :slight_smile:
As i keep pointing out TDF started up almost exactly 1 year ago.  All the other apps have a number of people that work on them or happily move between the different apps but none touch base.

Quirks and regressions are quite common in Base between one release of LO and another.  The regressions sometimes get posted as bug-reports but almost no devs are working on Base so they don't get fixed.

The current 'plan' of sit&wait has NOT worked during the last 1 year and shows no sign of working soon.

It would be nice to have a list dedicated to Base where we could discuss issues about how to re-invigorate that part of the project. 
Regards from
Tom :slight_smile:

Hi :slight_smile:
Hmmm, if you depend on Base then i think it's about time you start looking into switching to Kexi.  It supports a variety of back-ends, just as Base does, but it does have a large number of devs actively working on it.

I think we have to start recommending Kexi to anyone that has any problem with Base as Base's problems are unlikely to get fixed given the determination here of blocking any plans to develop a Base community within TDF. 
Regards from
Tom :slight_smile:

Hi Tom,

I am sorry but there is NO suitable list to discuss Base. Almost no-one that has expressed any interest in working on Base is on the list you discuss and they are not willing to join a high-traffic list that knows nothing about using Base.

this is the *wrong* list to discuss about base. *Please* move the discussion over to discuss@, as Thorsten asked.

Thanks,
Florian