Re: For all .NET developers: A new SpeechDispatcher client library

[Rastislav Kish]

Hello,

I was thinking about this before.


The thing is, I wouldn't be opposed from a conceptual point of view.

The important question is how to do it.


A problem I see with a complete transfer of the project under the
speechd repository is complicated management.

The library has its own versioning, its own development cycle.

In a language like C#, a higher-level abstraction will be necessary
sooner or later, not just to add things like blocking Speak or more
flexible priority setting scheme, but also because well, having a
Callback of type Action<CallbackType, string?>, when the second argument
exists solely for case when CallbackType is IndexMark and is null for
all other cases is...


I've choosen that approach because it was straight-forward and I did not
want to experiment much when the primary purpose was to get things working.


But now, when that is fulfilled, the time will come to think of
something better.


I'm yet to decide whether the change is going to be in the low-level
classes or it will be brought by the new abstraction, but in any case,
it will be necessary.


Thus, being hosted by a repository which is although related in topic,
but mostly unrelated in actual code, doesn't sound very well.


Of course, I would have my fork, I could even create local development
branches and tags, but then we'd practically have two histories of two
projects, with the same name on two different places.

Sounds unnecessarily messy and confusing.


Another approach would be to use git submodules.

This way, the library could live its life, while the speechd repository
could include it's releases according to stability requirements.


The only disadvantage of submodules is their complicated
synchronization, not just git doesn't clone them by default, but I'm not
sure whether pulls do update them, either.


So, the matter is kind of complicated. May be we should wait a bit for
the library to get tested, and include it afterwards, when all
significant work will be already done.


But when is that going to happen... I don't know.


Best regards


Rastislav


Dňa 8. 2. 2022 o 13:02 Samuel Thibault napísal(a):
> Hello,
>
> Rastislav Kish, le mar. 08 févr. 2022 10:20:16 +0000, a ecrit:
>> However, one particular drawback of using it on Linux was that there was
>> no library for using Speech dispatcher, i.e. getting speech was a bit
>> problematic.
>> That's why I've decided to develop a full-fledged Speech dispatcher in
>> pure C#, and as of today, Im happy to release it for anyone to use under
>> the terms of the GNU Lesser General Public License v2.1.
>> For more information, see the official project repository:
>> https://github.com/RastislavKish/SpeechDispatcher
> Could it makes sense to integrate it into the speechd repository, along
> the other bindings? That would make it much more apparent to people
> trying to use speech in C# applications.
>
>> I admit the documentation is kind of... non-existent at this point. I'm
>> not sure when do I get to writing one, I have lot of other projects
>> going on, so I don't want to spend much time on tasks like this.
> I'd still say that documentation if as important, if not more, as
> implementation.
>
> Samuel