[Emacspeak] Re: Mac server python version requirements, go version and intro

[T.V Raman <raman@xxxxxxxxxxx>]

SGTM on all fronts.

Again, to "ship with emacspeak"

Steps:

1. First get something working; if it works for you, it's paid for
   itself.
   2. Next, get it into Git and a state where I can accept it into the
      repo --   or if you prefer you can have your own repo, these are
      not  mutually exclusive.

      3. Once it's in Git, users get it when they clone the repo, they
         can build it themselves --  many users on Linux run from Git
         for instance because the distros are behind, and Mac : there
         is no distro;--)

         4. Go is a nice language I'v used it some, all you say about
            async, channels etc. is true. But all that said, you're
            still at the mercy of Apple's  audio APIs unless you plan
            to go all the way to the bare metal, which I suspect their
            platform wont let you do anyway.

           =20
Robert Melton writes:
 > Raman--
 >=20
 >=20
 > > 1. You should not let things like "build chains " etc guide your
 > >   exploration
 >=20
 > Great point!  I should have been more clear in my concerns.  My conc=
erns=20
 > were around getting the contribution into mainline Emacspeak.  I wan=
t to=20
 > make something that can be shipped "in the box" with Emacspeak.=20
 >=20
 >=20
 > > 2.  All the new language platforms, rust, go etc (heck even Haskel=
l)
 > >    have a large eco-system of libs and build chains...
 >=20
 > Haha, I agree, but I really want to lower the barrier of entry an in=
ch or=20
 > two if I can.  The usability gap between Emacspeak and anything else=
=20
 > I have tried in the last year is -- astonishing.  I really want to b=
ring more=20
 > low vision and blind users into the fold.  Using VSCode was extremel=
y=20
 > painful... I tried hooking Kakoune up to the mac server before I rea=
lly=20
 > understood anything, that was an abject failure.=20
 >=20
 >=20
 >=20
 > > 3. So by all means experiment with a server written in Go; though =
I fail
 > >   to understand why it would be performant
 >=20
 > I have been considering this a lot.  Specifically what causes the la=
g in the=20
 > python code.  I have yet to really trace it down, but it assuredly i=
s there,=20
 > when digging through lines with swiper (what a great plugin BTW) I c=
an=20
 > end up with a good few seconds of backlogged audio icons.  I am star=
ting
 > to think it has to do with the objective c delegate callbacks for co=
mpletions
 > stacking up and blocking (which even would stop processing on incomi=
ng=20
 > stdin till they unwind).=20
 >=20
 > Part of the case for Go can be made for async/await as well -- which=
 is just
 > the ability to do something else while waiting on IO.  However go ha=
s some
 > other real nice features for audio playback, soft-realtime GC, true =
threading
 > across multiple threads with no dreaded, yet unduly vilified, GIL.  =
The way
 > Go does interfaces well and cross platform support and building is a=
lso insanely=20
 > useful for something that might have a lot of pluggable TTS and audi=
o play
 > systems. Go is a good fit for this problem domain, great at doing lo=
ts of
 > things at once and reacting to stuff in flight.
 >=20
 > Honestly, I think the right answer is probably "why not both=3F"  I =
think I will
 > continue on the Python fixes for now -- as it is a fully functional =
server and=20
 > I have major things I do not understand about how the commands work =
as=20
 > of yet... is there a list of effects like echo and other stuff I cou=
ld reference=20
 > for example.
 >=20
 >=20
 > > P.S. You're a breath of fresh air -- really appreciate your openne=
ss to
 > > learning new things and for your fix to the Mac Server yesterday.
 >=20
 > Thanks!  Happy to be part community so open to contributions.  I hop=
e
 > to ship another fix and possibly a requirements.txt so people can in=
stall
 > the deps with one command... possibly even follow the lead of the ot=
her
 > servers "make mac" would validate and install what is required.
 >=20
 > --=20
 > Robert "robertmeta" Melton
--=20

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
=E2=99=89 Id: kg:/m/0285kf1  =F0=9F=A6=AE

-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮