if anybody manages to make ats2lua, that would be cool competition for
typed-lua
p.s. I would be really rather remiss to not call out Felix as well as
a language I have long lusted after. Whereâs enough free time for me
to get it installed, even?
Sure. This one is on my to-do-list.
I donât see atscc2lua as a competition for typed-lua. It is a tool for
writing quality lua code by taking advantage of the support for types
and templates in ATS.
In the history of mankind, is there such a language X for which typed-X
is more of a success than a failure?On Tuesday, December 16, 2014 7:37:58 PM UTC-5, Raoul Duke wrote:
if anybody manages to make ats2lua, that would be cool competition for
typed-lua
for me, non-typed is a non-starter. So while I have heard a lot about
Lua, I wasnât going to touch it. Ditto I donât like JavaScript much.
So I have tried typescript, but the compiler is too slow. (I have not
yet tried Flow myself, maybe thatâs not so bad.) Also, JS is sorta
hard to run fast, especially on constrained devices. Lua has embedded
Lua which might be nice. Thus I tried typedlua recently and was happy
about how fast the compiler is at least on stupid simple examples â
simple examples that failed to be fast with typescript, anyway. So in
my book, for my preferences and purposes, typed-X is nigh always
better.
(of course, I would love to also make use of ATS especially since the
set of back-ends has grown, because it could let me target C on the
low-end devices (or embedded Lua over to javascript for
desktop/web.)
and I should hope anybody doing safety critical stuff uses typed
languages in addition to all the intensive testing they should be
doing as well
I thought you would say Scheme/Racket(typed Scheme)
I myself once tried to add types to scheme (around 2002) but gave up
quickly.I took a quick tour over typed-lua. Of course, a BIG concern is that this
kind of
soft/gradual typing could lead to many false alarms or let type-errors slip
through.
This is not unlike the warnings issued by clang as we saw today.In any case, one needs to use it for a while in order to get a sense
whether it really
works. Or ask Scheme/Racket users?
I tried typed Racket for a while. It is nothing short of dreadful, I
think it only exist for the sole purpose of being able to say âhey, we
have static types too so you canât say XYZ is better due to having
them!â. Itâs constraining to use, annotations need to be scattered
everywhere, ton of stuff has fairly silly types just to accommodate for
LISPiness of the base language. Itâs not nice to use at all.
I agree with Hongwei Xi in that I personally donât know any typed-X
thatâs more successful than just X: you get all the faults of X with all
the faults of a typing system that was just tacked on as an after-thought.
Personally I recommend using a language designed around typing to begin
with such as Haskell. If youâre doing safety critical, typed-X seems
like a pretty terrible choice: maybe better than just X but certainly
not nicer to write in.> On Wed, Dec 17, 2014 at 2:31 AM, Raoul Duke rao...@gmail.com wrote:
for me, non-typed is a non-starter. So while I have heard a lot about
Lua, I wasnât going to touch it. Ditto I donât like JavaScript much.
So I have tried typescript, but the compiler is too slow. (I have not
yet tried Flow myself, maybe thatâs not so bad.) Also, JS is sorta
hard to run fast, especially on constrained devices. Lua has embedded
Lua which might be nice. Thus I tried typedlua recently and was happy
about how fast the compiler is at least on stupid simple examples â
simple examples that failed to be fast with typescript, anyway. So in
my book, for my preferences and purposes, typed-X is nigh always
better.(of course, I would love to also make use of ATS especially since the
set of back-ends has grown, because it could let me target C on the
low-end devices (or embedded Luaover to javascript for
desktop/web.)and I should hope anybody doing safety critical stuff uses typed
languages in addition to all the intensive testing they should be
doing as wellâ
You received this message because you are subscribed to the Google Groups
âats-lang-usersâ group.
To unsubscribe from this group and stop receiving emails from it, send an
email to ats-lang-user...@googlegroups.com.
To post to this group, send email to ats-lan...@googlegroups.com.
Visit this group at http://groups.google.com/group/ats-lang-users.
To view this discussion on the web visit
https://groups.google.com/d/msgid/ats-lang-users/CAJ7XQb4xG5rGGnoJ5mCt_Rp6Hwfi2RChY06fhMD2kcJQHLczjQ%40mail.gmail.com
.
Mateusz K.
hi,
I always ethically liked the idea of TR. I havenât used it. I read the
Racket mailing list so Iâve seen discussions about tricky stuff with
it. Thanks for the real-world blunt report on its usability! so sad.
Some people I know personally and respect a lot have been Racketeers
tho they werenât the TR type of person.
Heck, Stappl was done with Racket. So yeah I guess Racket >> TR apparently?
I want to love Bigloo but recently it wouldnât install easily. I want
to love Chickenâs optional typing, but I worry that thereâs no
interactive debugger story there. Iâd go for type hints in CL but CL
probably wonât run games well on my devices - tho I wish I had the
time to try ECL. Or the money to pay for mocl. Shen is super cool to
me, but whereâs the super good debugger?
I appreciate the guidance of going for a language that is typed out of
the gate. Probably I do like the MLs the best in general. But
languages like that are generally hard to get on iOS & Android for
doing apps there, let alone on smaller old retro video game hardware
it seems.
Thatâs one reason ATS appeals to me so much; I can use C which can go
smaller places than most languages. And I can use standard C tools to
study the program. But I am worried that in reality the generated code
is too tricky to really understand while stepping through in GDB.
(Many other super cool languages have this issue. Probably more the
rule than the exception. Shen, Haxe, CPC, et. al. ⌠and Scheme, ML,
Haskell (tho Neil Mitchell is a stud and did work on debuggers for
Haskell) et. al.)
But then it looks like eLua can, too, fit on smaller devices â albeit
probably a lot slower? So Iâm trying to learn and get set up with Lua
for a while now for whatever it is worth! typedlua for the win i hope.
What do people think of Dialyzer? I guess Iâve mostly only heard good
things about it. What about Flow for JS? Typescriptâs compiler is just
too slow to be an even remotely realistic thing in my mind. Like
Scalaâs.
Haxe uses Ocaml and compiles really fast even on smaller older lamer
laptops. But I have had sad experiences with the ecosystem there. I
never did find out why the Android version of my game steadily leaked
something-or-other whereas the iOS one did not. Talk about leaky
abstractions.
sincerely.
depends on how you define success vs. failure.
Yes, it is totally subjective.
If you give us the name of a language that has the potential to be
X (for which a typed X is more of a success than X), then we can
start to evaluate it :)On Wed, Dec 17, 2014 at 2:14 AM, Raoul Duke rao...@gmail.com wrote:
In the history of mankind, is there such a language X for which typed-X
is more of a success than a failure?ouch.
depends on how you define success vs. failure. having lame crash and
burn bugs, lame security holes, difficult maintenance which thus costs
more $ in the long run, etc. are all in my mind valid reasons to say
that, yes, typed versions are better!!!â
You received this message because you are subscribed to the Google Groups
âats-lang-usersâ group.
To unsubscribe from this group and stop receiving emails from it, send an
email to ats-lang-user...@googlegroups.com.
To post to this group, send email to ats-lan...@googlegroups.com.
Visit this group at http://groups.google.com/group/ats-lang-users.
To view this discussion on the web visit
https://groups.google.com/d/msgid/ats-lang-users/CAJ7XQb47HvcHAZPP_Lx%3Duu_PnVWGAOnTOJnKEV6R0hZbAsyXJA%40mail.gmail.com
.
By the way, part of the reason is that lua-jit gets close to C in some
benchmarks. (according to some people who heard it from somebody else
who maybe read it online, once.)
I thought you would say Scheme/Racket(typed Scheme)
I myself once tried to add types to scheme (around 2002) but gave up
quickly.
I took a quick tour over typed-lua. Of course, a BIG concern is that this
kind of
soft/gradual typing could lead to many false alarms or let type-errors slip
through.
This is not unlike the warnings issued by clang as we saw today.
In any case, one needs to use it for a while in order to get a sense
whether it really
works. Or ask Scheme/Racket users?On Wed, Dec 17, 2014 at 2:31 AM, Raoul Duke rao...@gmail.com wrote:
for me, non-typed is a non-starter. So while I have heard a lot about
Lua, I wasnât going to touch it. Ditto I donât like JavaScript much.
So I have tried typescript, but the compiler is too slow. (I have not
yet tried Flow myself, maybe thatâs not so bad.) Also, JS is sorta
hard to run fast, especially on constrained devices. Lua has embedded
Lua which might be nice. Thus I tried typedlua recently and was happy
about how fast the compiler is at least on stupid simple examples â
simple examples that failed to be fast with typescript, anyway. So in
my book, for my preferences and purposes, typed-X is nigh always
better.(of course, I would love to also make use of ATS especially since the
set of back-ends has grown, because it could let me target C on the
low-end devices (or embedded Luaover to javascript for
desktop/web.)and I should hope anybody doing safety critical stuff uses typed
languages in addition to all the intensive testing they should be
doing as wellâ
You received this message because you are subscribed to the Google Groups
âats-lang-usersâ group.
To unsubscribe from this group and stop receiving emails from it, send an
email to ats-lang-user...@googlegroups.com.
To post to this group, send email to ats-lan...@googlegroups.com.
Visit this group at http://groups.google.com/group/ats-lang-users.
To view this discussion on the web visit
https://groups.google.com/d/msgid/ats-lang-users/CAJ7XQb4xG5rGGnoJ5mCt_Rp6Hwfi2RChY06fhMD2kcJQHLczjQ%40mail.gmail.com
.
In the history of mankind, is there such a language X for which typed-X
is more of a success than a failure?
ouch.
depends on how you define success vs. failure. having lame crash and
burn bugs, lame security holes, difficult maintenance which thus costs
more $ in the long run, etc. are all in my mind valid reasons to say
that, yes, typed versions are better!!!