minutes of ESC call ...

classic Classic list List threaded Threaded
25 messages Options
12 « Prev
Tor Lillqvist-2 Tor Lillqvist-2
Reply | Threaded
Open this post in threaded view
|

Re: [Libreoffice-qa] minutes of ESC call ...

I have an even blunter suggestion: In a branch (branched off 5.2, for instance), we unpublish and rename all existing API. (For instance, s/com::sun::star/libreoffice/.) We make this very public, and make a test build available that contains it. And then we wait for reports what external software actually breaks against this test build. As such reports come in, we rename back those APIs, and mark them published. Once a week we distribute a new build from that branch. If after half a year no new complaints arrive, we know what APIs actively used and/or developed external software actually uses, and do the corresponding change also in master.

Of course I don't really believe this suggestion has any chance of being accepted. But it is Friday, and one can dream.

--tml


_______________________________________________
LibreOffice mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/libreoffice
Michael Meeks-5 Michael Meeks-5
Reply | Threaded
Open this post in threaded view
|

Re: [Libreoffice-qa] minutes of ESC call ...

In reply to this post by Bjoern Michaelsen
Hi Bjoern,

On 16/12/16 12:11, Bjoern Michaelsen wrote:
> Maybe we should change the meaning of "published" then? Currently, "published"
> means a onesided promise to each and eevery person on this planet -- people that we have
> no relation with at all -- that we will never ever change this API and we dont
> ever get anything in return.

        Fair enough.

> 1/ We unpublish all API
> 2/ We give UNO user an opportunity to ask for republishing specific parts of
>    the API, when they provide a reasonable use case and promise to be the
>    "client steward" for these.

        Sounds sensible to me; ideally with unit tests etc. FWIW - I'd suggest
that much of the udkapi can remain published though.

> 3/ We continue to change unpublished API as core developers see reasonable. We
>    promise to release changes to these newly republished APIs only after
>    checking back with the "client steward" for that part of the UNO API for
>    advisory _non-blocking_ input[1].

        I like a mapping of real people to bits of API. Particularly if we
focus just on those bindings that have real problems with extension
(currently) - like Java, C++ etc.

> I see multiple advantages to this:
>
> - we (core devs) still retain the perogatibve to do the ultimate decisions on
>   UNO API changes

        You know - I still really like having some area where people can work
without asking someone; I think having an "ultimate decision making
authority" that is frequently wielded when necessary is a net negative
in general =) It is also just generally a social ill to have to have
guardians watching left & right and jumping on people - the less we can
do of that the better.

> There might be some hope that UNO API users like WollMux, Mendeley, Zotero
> might be interested in this -- and by talking to them instead of with
> $ANONYMOUS_GUY_ON_THE_INTERTUBES we might get a sensible feedback channel and
> bring out ecosystem closer together -- as they have incentives to join this
> discussion.

        I would love some reliable way whereby: without asking them we can get
a dump of all the interfaces they use =) I'm well up for making eg. our
external C++ bindings less efficient - if we win that =) IMHO making
good, data-driven decisions, and even better relational decisions [ we
can go talk to the one user of foo Interface ;-] is -far- more useful,
the Mono guys had a tool to help with this. I love your idea of getting
the extension guys closer to us and getting some database of people who
care about interfaces.

        ATB,

                Michael.

--
[hidden email] <><, Pseudo Engineer, itinerant idiot
_______________________________________________
LibreOffice mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/libreoffice
Bjoern Michaelsen Bjoern Michaelsen
Reply | Threaded
Open this post in threaded view
|

Re: [Libreoffice-qa] minutes of ESC call ...

Hi,

On Fri, Dec 16, 2016 at 12:24:33PM +0000, Michael Meeks wrote:
> > - we (core devs) still retain the perogatibve to do the ultimate decisions on
> >   UNO API changes
>
> You know - I still really like having some area where people can work
> without asking someone; I think having an "ultimate decision making
> authority" that is frequently wielded when necessary is a net negative
> in general =) It is also just generally a social ill to have to have
> guardians watching left & right and jumping on people - the less we can
> do of that the better.

That is what I am saying though? I do NOT want to change much about where the
decisions are made (core devs and ESC). But these feedback channels might
provide us with a more educated base for decision. I assume in the end we will
have:

- 80% changes are uncontroversial anyway
- 10% find initial discussion, but are accepted once rationale is discussed
  (plus those caring about this are informed earlier an can adapt in the time
  to release -- also learning more about core development)
-  5% find discussion with some clients still disagreeing with core devs. Though
  luck, but no change to the situation before for the most part. Also adaption
  on the client side is still possible and we save ourselves from "all rants
  on release date" phenomena
-  5% core devs learn about a use case they didnt consider and come back with a
  better suited change

> I would love some reliable way whereby: without asking them we can get
> a dump of all the interfaces they use =) I'm well up for making eg. our
> external C++ bindings less efficient - if we win that =) IMHO making
> good, data-driven decisions, and even better relational decisions [ we
> can go talk to the one user of foo Interface ;-] is -far- more useful,
> the Mono guys had a tool to help with this. I love your idea of getting
> the extension guys closer to us and getting some database of people who
> care about interfaces.

Also a nice idea. Im unsure how doable automatically that is with the
abstraction layercake of UNO, but indeed would be interesting. FWIW, I already
wrote two draft ideas for the budget from this:
 https://wiki.documentfoundation.org/index.php?title=Development%2FBudget2017&diff=140531&oldid=140527

Best,

Bjoern
_______________________________________________
LibreOffice mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/libreoffice
Thorsten Behrens-6 Thorsten Behrens-6
Reply | Threaded
Open this post in threaded view
|

Re: [Libreoffice-qa] minutes of ESC call ...

In reply to this post by Bjoern Michaelsen
Hi Björn,

Bjoern Michaelsen wrote:
> 1/ We unpublish all API
>
Ok so far -

> 2/ We give UNO user an opportunity to ask for republishing specific parts of
>    the API, when they provide a reasonable use case and promise to be the
>    "client steward" for these.
>
This is where the plan breaks already.

To relay some stories from real life - there's _so_ much integration
with especially OpenOffice out there, coming from legacy systems, tons
of macro-ized applications, sometimes bespoke development with the
company simply selling, not developing anymore - that by the time
those deployments 'see' our unpublished API (and subsequent changes
thereof), it's gonna be much, much, much too late to do anything about
it.

Those people will then simply walk away.

The idea that we could somehow entice a sizeable chunk of our API
consumers to be part of our community is a pipe dream. Like it or not,
with taking over market share from OpenOffice we inherited that API
promise, and I'd really prefer to continue serving that market
segment.

I mean, Linux has not been exactly timid in changing implementation
quite rapidly, still they managed to keep the userland API reasonably
stable. Why can't we? Except for the repulsiveness of XFoo2, XFoo3,
and some extra day of hacking here or there - did this really hold us
back so far?

> There might be some hope that UNO API users like WollMux, Mendeley, Zotero
> might be interested in this -- and by talking to them instead of with
> $ANONYMOUS_GUY_ON_THE_INTERTUBES we might get a sensible feedback channel and
> bring out ecosystem closer together -- as they have incentives to join this
> discussion.
>
Orthogonally to the above, that's a really worthwhile idea. I was once
already kicking the idea of an API forum / mailing list / whatever
around; perhaps -- if properly seeded & tended -- the dev askbot
instance could be that?

Cheers,

-- Thorsten

_______________________________________________
LibreOffice mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/libreoffice

signature.asc (968 bytes) Download Attachment
Bjoern Michaelsen Bjoern Michaelsen
Reply | Threaded
Open this post in threaded view
|

Re: [Libreoffice-qa] minutes of ESC call ...

Hi,

On Fri, Dec 16, 2016 at 11:40:21PM +0100, Thorsten Behrens wrote:
> To relay some stories from real life - there's _so_ much integration
> with especially OpenOffice out there, coming from legacy systems, tons
> of macro-ized applications, sometimes bespoke development with the
> company simply selling, not developing anymore - that by the time
> those deployments 'see' our unpublished API (and subsequent changes
> thereof), it's gonna be much, much, much too late to do anything about
> it.

As said in my reply to Michael already, this has little to do with the actual
decision making except for us to get a better information base.

Your statement and Michaels statement display the antipodes of these arguments
and I assume we will continue to have these discussion on the ESC on a
case-by-case base -- possibly just somewhat better informed now and then.

> Orthogonally to the above, that's a really worthwhile idea. I was once
> already kicking the idea of an API forum / mailing list / whatever
> around; perhaps -- if properly seeded & tended -- the dev askbot
> instance could be that?

see draft at: https://wiki.documentfoundation.org/index.php?title=Development%2FBudget2017&diff=140532&oldid=140527
(also mentions askbot already)

Best,

Bjoern
_______________________________________________
LibreOffice mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/libreoffice
12 « Prev