[Opentts-dev] merging opentts and speechd was Re: Fwd: Re: OpenTTS 0.1 released.

[Aleksandar Kostadinov]

Hello Jan,

I see you care a lot about these matters but some of your assumptions
are not correct. Most importantly the role of volunteer contributors.
How you describe it it is not possible to create a good opensource
ecosystem around SD because volunteers who are often most motivated
are not held on an equal ground.

Look at Linux. Contributors are not asked if they are payed to work on
the kernel or not. Which company hires them and such stuff. I would
highly recommend you to search in google for the linux development
model and reconsider your principles.

I'm not a contributor and am actually subscribed to the list somehow
by luck (good or bad, I don't know) but I think if you had somebody to
just review the patches and direct volunteers at how to fix them to be
acceptable, that wouldn't cost you much and project wouldn't have been
in such a situation.
I feel that opentts guys (although I haven't talked to them) would
accept Brailcom's leadership if clear rules are agreed upon for code
quality, commit rights and so on. For example you can think of
something like commit on 3 acks from established developers or some
other scheme.

See you are writing the code for the people and as I see it, mostly
people that use the code are willing to contribute code to it. So you
need to be listening to them and understand what do they want and what
do they need rather than think your opinion is the best. It's all
about trust that needs to be established and then things go smoother.

IMHO you are proposing an approach that just cannot scale. Excuse me
but your promise sounds very vague and somehow leaves volunteer
contributors in the air as they were before.
By using a more opened and scalable approach, the future of SD will
not be determined only from the funds sent to your organization but
will be much more stable and predictable. As many projects start to
rely on SD, you will see more and more developers contributing (many
of them payed by other companies).

I see SD your baby but it has grown and same approaches just don't
work (proven by the created fork).

It's all IMHO though...

On Wed, Jun 9, 2010 at 4:14 PM, Jan Buchal <buchal at brailcom.org> wrote:
> Hello,
>
> let me express my opinion of both the past and the current events regarding
> the Speech Dispatcher project and the newly created OpenTTS project. I am
> the founder and long-term director of the Brailcom company, a company
> founded to help the visually-impaired. The nature of the company is
> non-profit, even though I am convinced that is not important. What is
> important is that our company employs people who create and work on
> projects that they believe can substantially help the visually-impaired.
>
> Someone can also find it important that I am myself fully blind and am
> involved in this for nearly 20 years. Brailcom, as I have already mentioned
> in another letter, is a professional organization; that means its staff are
> paid full-time employees. That does not mean that I don't want to
> collaborate with volunteers, but rather that I consider it necessary and
> obvious that everyone is appropriately rewarded for their work. We do not
> live in a society without money and money is, as you know, important for
> our lives. Of course I highly value everyone who devotes time to the
> others, however that does not mean that someone who is paid would be
> helping less than others, or that they would not deserve my admiration and
> respect. The only thing that decides is the quality work and whether the
> work is really beneficial.
>
> I am convinced that long-term projects that require highly-qualified
> manpower have to be financed in their basis, that those who are core
> members of the of the development team must be sufficiently materially
> secured so that they can work on the project on a long-term basis.
> Volunteers are and can be a substantial help, but they rather suppport and
> supplement the core developers of the project. Volunteer also has to make a
> living and if their life situation changes, then they often no longer have
> much time to support a specific project. In order to preserve continuity
> and quality of any project, it is necessary that the core developers are
> not, if possible, changed, so that the ideas that seem simple in the
> beginning, can really be realized. To achieve that, a lot of time and deep
> experience is needed.
>
> If I have not said anything new so far, or if someone is bored by my
> thoughts, try to continue in reading so that you possibly gain the big
> picture of the issue. That means how the development is organized, or what
> is the process of developing projects in our company.
>
> Every project, including Speech Dispatcher, has its responsible person,
> a.k.a. maintainer. Apart from this person, so called opponents are involved
> in the project who supply direct feedback to the person working on the
> problem. Opponent does not do testing, but real opponency, where one
> developer investigates and contemplates proposals or implementation of the
> other developer. Internal testing of course takes place as well, before we
> invite broader public to testing.
>
> As has been already said, Speech Dispatcher is and continues to be one of
> the key projects of Brailcom. The origins of Speech Dispatcher reach
> sometimes into year 1997, even though its current implementation is dated
> approximately 10 years ago. During the whole development, it was sometimes
> easy and sometimes hard to secure the needed funds from various grants and
> dotations. Because I however want that experienced and capable developers
> stay in our company, I have to manage to offer them appropriate financial
> backing. They of course are not doing the work because of money, but there
> is no doubt they deserve the reward. Not only because we are successful at
> securing financial resources is the Brailcom team stable and many of its
> members have been in Brailcom for nearly as long as is the history of
> Speech Dispatcher.
>
> It is however not always possible to secure finances specifically for the
> development of Speech Dispatcher. It is often hard to explain to the public
> officers of the state that Speech Dispatcher is something that can help the
> visually-impaired. Brailcom is developing various other projects like
> electronic language textbok, digital language school, or chemistry for the
> visually-impaired -- those are not directly related to Speech Dispatcher,
> but in their budgets, they were or are partially supporting the development
> of Speech Dispatcher.
>
> As I have already said, Speech Dispatcher has been with us for more than 10
> years and also much sooner than for example gnome speech. We have attempted
> for a very long time for Speech Dispatcher to become the default speech
> interface for the Free Accessibility Infrastructure and in this sense I
> have conducted numerous talks with the managers of the company SUN. I hoped
> that it will be recognized that this component of accessibility
> infrastructure, with its architecture, is very important and that funding
> for its support will be found. So far nothing has changed in this regard
> and Speech Dispatcher is financed solely from the funds of Brailcom. Last
> time I made this offer was in bugzilla where the issue about replacing
> gnome speech was being discussed. However I also mentioned there that we
> are searching for further funds so that we can continue the development
> with more intensity. It did not work out as fast as we wished. But we have
> some funds available right now so we could intensify the development of
> Speech Dispatcher. I can assure everyone that the efforts of volunteers
> around the OpenTTS project did not speed up anything in this matter. The
> project from which the development of Speech Dispatcher is currently being
> financed was approved in the end of December last year, we just had to wait
> for signing the contract with the Government of the Czech Republic and then
> of course wait for the financial funds themselves. That the renewal of work
> on Speech Dispatcher happened at the same time as the activities of people
> around OpenTTS is a coincidence. It is not correct to assume different
> conclusions.
>
> If it is pointed out that Brailcom did nothing during the last year and a
> half, that it was impossible to continue in the development, it is
> necessary to state that such conclusions are not based on proper analysis
> of the situation. Yes, Brailcom did not have the needed funds available,
> but it was open to cooperation with volunteers, which naturally does not
> mean that we resigned from leading the project. I think no obstacle was in
> the way of tighter cooperation and the same is true also now.
>
> I understand the current state as that there is a distrust towards the
> ability of Brailcom to successfully manage the project, that maybe there is
> generally a distrust that companies can manage the needed support, or
> rather the belief that a group of volunteers is more reliable because they
> are not dependent on financial resources. I cannot agree with that. As I
> have mentioned, I am convinced that such serious and important projects
> like projects from the area of Free Accessibility require permanent
> financial support, and that without this support it is possible to achieve
> something but it is not possible to bring a real alternative to proprietary
> applications and keep pace with the development of IT.
>
> It was being pointed out with respect to us that lately, that means before
> the entry of the OpenTTS group into the development of Speech Dispatcher,
> that Speech Dispatcher was unstable and unreliable. Also in this case it is
> necessary to analyze the situation properly and only after that come to
> proper conclusions. As many of you know, the big revolution for everyone
> needing a sound output was the deployment of the PulseAudio system. That
> however introduced many problems. As soon as PulseAudio appeared, we began
> to intensively communicate with Mr. Poetering, the main PulseAudio
> developer, and discuss how Speech Dispatcher should use PulseAudio. That
> meant hundreds of hours of work and many changes even in PulseAudio itself.
> PulseAudio itself was however being developed rapidly, so it was not
> possible to find a stable solution quickly, as we wished. Now, thanks also
> to volunteers, we are in a state where communication with PulseAudio partly
> works. Current state is however still provisional and should be from the
> long-term perspective understood as just a workaround. Someone could
> challenge this with the fact that similar instability manifested itself
> with ALSA. Yes, but why? Many people were seeking answer and even today
> some questions remain unanswered. It is again necessary to take into
> account that Speech Dispatcher was often using a very unstable sound system
> and sound card drivers developed with the reverse engineering method.
>
> Back to that distrust towards Brailcom, the distrust of whether the
> Brailcom team is capable to run such an important project as Speech
> Dispatcher. Firstly I am very pleased that Speech Dispatcher is seen by
> many as an important project and that they overtook care of it, even though
> I think it would be better that the fork did not take place. I am convinced
> that Brailcom always met its commitments and proved that is has a team of
> qualified developers that are capable to lead such an important project. Of
> course I cannot manage to convince others to think the same. I just ask you
> to draw conclusions from a real, proper analysis of the situation so that
> we do not unnecessarily waste the already limited resources that can be
> used for those that need our help.
>
> I therefore propose to those who decided to fork Speech Dispatcher to
> consider whether this way, that is the way of fork, is really more
> beneficial than the previous model of development, to consider whether it
> is possible to have faith in the leadership of Brailcom and the developers
> of Brailcom and to consider whether it is possible to accept the offer that
> I made when I was recently asked to state how much time can we devote to
> the development of Speech Dispatcher. I can't under pressure promise
> something that we cannot fulfill. I can however promise that we will pursue
> the project if there is a real interest in Speech Dispatcher, that we
> currently have funds to lead the project, and that we will have the funds
> also in near future. Everything however depends on many circumstances that
> are unpredictable, so it would not be serious to claim that it will not be
> a problem to secure a stable amount of workhours. There are many problems
> in the area of Free Accessibility that have to be solved and sometimes it
> seems like we are trying to extinguishing several fires at the same time.
> Despite that, I am an optimist and I believe that many things can change,
> for example by finding a common way with those that feel they will do the
> Speech Dispatcher project better good by forking it.
>
> To finally summarize the current state and restate the offer, I say that we
> currently have funds that sufficiently secure the subsequent development of
> Speech Dispatcher and that if we are joined by skilful and willing
> volunteers who are willing to respect Brailcom's way of development, then
> we can significantly advance the project. It is also possible that after
> some time I will offer some volunteers a way of paid cooperation. I am not
> saying this to "bribe" someone, but because I do not find it natural that
> someone works for the others without being paid.
>
> With friendly regards,
> Jan Buchal -- CEO
> Brailcom, o.p.s.
>
> _______________________________________________
> Speechd mailing list
> Speechd at lists.freebsoft.org
> http://lists.freebsoft.org/mailman/listinfo/speechd
>