• Leandro Lucarella (16/16) Jan 11 2007 Hi, I don't want to be rude, *really*, but why don't you release Tango?
• Lars Ivar Igesund (27/42) Jan 11 2007 The first release will be this alpha relase you talk of. As for not
• Leandro Lucarella (17/33) Jan 11 2007 But you are thinking only about users, what about developers? I think
• Sean Kelly (7/23) Jan 11 2007 The problem is that once Tango is available to the public there's no way...
• zz (4/12) Jan 11 2007 Take your time it's allways better to have great documentation from the
• John Reimer (17/32) Jan 11 2007 The argument could be made both ways, of course.
• Daniel Keep (9/9) Jan 14 2007 Ok, even if you're not ready to release just yet, can you at least let
• Lars Ivar Igesund (25/39) Jan 14 2007 From the first announcement:
• Daniel Keep (9/46) Jan 14 2007 *kack* *hiss* (don't mind me :P)

Leandro Lucarella <llucarella integratech.com.ar> writes:

Hi, I don't want to be rude, *really*, but why don't you release Tango? 
I don't care if it's not done yet, you can also make an alpha release 
and accept no patches/opinions of improvements.

I'm starting to find very annoying to read "when Tango is released", 
"Tango has this", etc. Show me the code! =)

I really think you are underestimating the D community if you think you 
could not get any good criticism/suggestions/patches before Tango it's 
all done.

I don't really get all the secrecy on Tango, I find it 
counterproductive, and please believe me, this is not trolling. I 
really, *really* think you are doing more harm than good to Tango 
keeping it secret.

-- 
Leandro Lucarella
Integratech S.A.
4571-5252

Lars Ivar Igesund <larsivar igesund.net> writes:

Leandro Lucarella wrote:

 Hi, I don't want to be rude, *really*, but why don't you release Tango?
 I don't care if it's not done yet, you can also make an alpha release
 and accept no patches/opinions of improvements.

The first release will be this alpha relase you talk of. As for not
accepting/looking at/considering patches/opinions etc, how could that help?
We certainly don't want to ignore anyone having something to contribute,
whether it is opinions, fixes or anything larger, but we need to feel
confident that we can actually handle incoming information.

Further on, Tango has been equally much about documentation, as about code.
Tango has very much code, but it hasn't had proper documentation besides
the API-docs. It will be hard to get proper feedback without proper
documentation to explain the concepts, especially now with so many new
users after the D 1.0 release. We need to get the documentation in a state
where it can present users with useful information.

 
 I'm starting to find very annoying to read "when Tango is released",
 "Tango has this", etc. Show me the code! =)

Tango is a software product, and as such it is delayed ;) As for "Tango has
this", sorry, but as we saw that a relase ready for D 1.0 slipped by, we
still wanted to tell the community that we are working on something cool.

 
 I really think you are underestimating the D community if you think you
 could not get any good criticism/suggestions/patches before Tango it's
 all done.

Oh, we're not underestimating the community at all, after all quite a big
part of it is already involved. I do think that heavy usage from the full
community at the stage we've been at lately (heavy refactoring and sorting
of minor issues), would only cause frustrations for both users and us. We
are doing this in our spare time, and we want to come to a level, where we
can actually handle the feedback.

 
 I don't really get all the secrecy on Tango, I find it
 counterproductive, and please believe me, this is not trolling. I
 really, *really* think you are doing more harm than good to Tango
 keeping it secret.

We think opposite :)

-- 
Lars Ivar Igesund
blog at http://larsivi.net
DSource & #D: larsivi
Dancing the Tango

Leandro Lucarella <llucarella integratech.com.ar> writes:

 I really think you are underestimating the D community if you think you
 could not get any good criticism/suggestions/patches before Tango it's
 all done.

 
 Oh, we're not underestimating the community at all, after all quite a big
 part of it is already involved. I do think that heavy usage from the full
 community at the stage we've been at lately (heavy refactoring and sorting
 of minor issues), would only cause frustrations for both users and us. We
 are doing this in our spare time, and we want to come to a level, where we
 can actually handle the feedback.

But you are thinking only about users, what about developers? I think 
developers are smart enough to know that some library is still 
experimental and take a look at it even just for curiosity.

I don't know, there so many examples of open projects that was developed 
this way and was successful. I think is all about the development model, 
the good old The Cathedral and the Bazaar[1]. I think DMD, even having a 
closed source backend, has many things of the bazaar model ("release 
early, release often" =).

Anyways is your software, is your decision =)

 I don't really get all the secrecy on Tango, I find it
 counterproductive, and please believe me, this is not trolling. I
 really, *really* think you are doing more harm than good to Tango
 keeping it secret.

 
 We think opposite :)

I hope I'm the one who is wrong =)

[1] I think everybody who is developing Tango has already read this, but 
just in case: 
http://catb.org/~esr/writings/cathedral-bazaar/cathedral-bazaar/

-- 
Leandro Lucarella
Integratech S.A.
4571-5252

Sean Kelly <sean f4.ca> writes:

Leandro Lucarella wrote:
 I really think you are underestimating the D community if you think you
 could not get any good criticism/suggestions/patches before Tango it's
 all done.

 Oh, we're not underestimating the community at all, after all quite a big
 part of it is already involved. I do think that heavy usage from the full
 community at the stage we've been at lately (heavy refactoring and 
 sorting
 of minor issues), would only cause frustrations for both users and us. We
 are doing this in our spare time, and we want to come to a level, 
 where we
 can actually handle the feedback.

 
 But you are thinking only about users, what about developers? I think 
 developers are smart enough to know that some library is still 
 experimental and take a look at it even just for curiosity.

The problem is that once Tango is available to the public there's no way 
to distinguish between "developers" and "users" (which I might describe 
as "experienced users" and "novice users").  Until we're ready, I really 
think it's best to limit availability to those testing and documenting 
the library.


Sean

zz <zz zz.com> writes:

Sean Kelly wrote:
 The problem is that once Tango is available to the public there's no way 
 to distinguish between "developers" and "users" (which I might describe 
 as "experienced users" and "novice users").  Until we're ready, I really 
 think it's best to limit availability to those testing and documenting 
 the library.
 
 
 Sean

Take your time it's allways better to have great documentation from the 
start. I know it will be great once released even though I can hardly wait.

Zz.

John Reimer <terminal.node gmail.com> writes:

On Thu, 11 Jan 2007 12:08:12 -0300, Leandro Lucarella wrote:

 Hi, I don't want to be rude, *really*, but why don't you release Tango? 
 I don't care if it's not done yet, you can also make an alpha release 
 and accept no patches/opinions of improvements.
 
 I'm starting to find very annoying to read "when Tango is released", 
 "Tango has this", etc. Show me the code! =)
 
 I really think you are underestimating the D community if you think you 
 could not get any good criticism/suggestions/patches before Tango it's 
 all done.
 
 I don't really get all the secrecy on Tango, I find it 
 counterproductive, and please believe me, this is not trolling. I 
 really, *really* think you are doing more harm than good to Tango 
 keeping it secret.

The argument could be made both ways, of course.

The D community will not be the only ones that judge Tango if it is
released before it is ready.  Those on the outside will be much more
critical of an incomplete library.  

Mostly the documentation has to be completed now, so that people will know
how to use it when its released. 

The "secrecy" is an attempt at being level-headed and conscientious. The
Tango team wants to see Tango at strong beta quality with a sufficient
level of core documentation before it's released. I think that's a fair
goal.  Otherwise, if it were released early, can you imagine the huge
technical support burden on the Tango team as a result of users not
understanding how to get the library working?

I think much of that can be avoided with good solid documentation from the
beginning, and I think that's what the Tango team is shooting for. 
Believe me, it's all for the best. :D

-JJR

Daniel Keep <daniel.keep+lists gmail.com> writes:

Ok, even if you're not ready to release just yet, can you at least let 
us know what's in Tango?

Over the past few days, I keep thinking "well, I need a library for X. 
I could implement it myself, but what if Tango has it, and then I have 
to go change all my code?".

Even just a quick run-down would be good.

	-- Daniel

P.S.  Keep up the good work; from the names that have been popping up in 
relation to Tango, I can't wait to see what it looks like :)

Lars Ivar Igesund <larsivar igesund.net> writes:

Daniel Keep wrote:

 
 Ok, even if you're not ready to release just yet, can you at least let
 us know what's in Tango?
 
 Over the past few days, I keep thinking "well, I need a library for X.
 I could implement it myself, but what if Tango has it, and then I have
 to go change all my code?".
 
 Even just a quick run-down would be good.
 
 -- Daniel
 
 P.S.  Keep up the good work; from the names that have been popping up in
 relation to Tango, I can't wait to see what it looks like :)

From the first announcement:

* Atomic mark/sweep garbage collection by default, with a malloc-based 
stub allocator to serve as an example for future development.
* User interception of important language and system-level errors and 
events.
* Various levels of concurrency, including process control, kernel 
threads, and fibers (stack threads).
* A robust IO framework based on the Mango design.
* An array of mathematics routines ranging from low-level IEEE interop. 
to high-level statistics and numerics.
* A networking model that is growing to encompass some of the most 
common IO designs and network protocols.
* Standard C, POSIX, and system API support.
* Basic text processing, a container package, logging facilities, and more.

I think you will find few surprises beyond this, but I can mention
non-blocking IO (selectors), ftp, http, locale, time and calendar support,
some crypto algorithms. The rest is in the details, I think, and what we're
still undecided on. There is a bit of useful template functionality in
there too.

-- 
Lars Ivar Igesund
blog at http://larsivi.net
DSource & #D: larsivi
Dancing the Tango

Daniel Keep <daniel.keep+lists gmail.com> writes:

Lars Ivar Igesund wrote:
 Daniel Keep wrote:
 
 
Ok, even if you're not ready to release just yet, can you at least let
us know what's in Tango?

Over the past few days, I keep thinking "well, I need a library for X.
I could implement it myself, but what if Tango has it, and then I have
to go change all my code?".

Even just a quick run-down would be good.

-- Daniel

P.S.  Keep up the good work; from the names that have been popping up in
relation to Tango, I can't wait to see what it looks like :)

 
 
 From the first announcement:

Oops.  Sorry :)

 * Atomic mark/sweep garbage collection by default, with a malloc-based 
 stub allocator to serve as an example for future development.
 * User interception of important language and system-level errors and 
 events.
 * Various levels of concurrency, including process control, kernel 
 threads, and fibers (stack threads).
 * A robust IO framework based on the Mango design.

*kack* *hiss* (don't mind me :P)

 * An array of mathematics routines ranging from low-level IEEE interop. 
 to high-level statistics and numerics.
 * A networking model that is growing to encompass some of the most 
 common IO designs and network protocols.

Now *that* sounds interesting... well, ok; pretty much all of the above 
sounds interesting, but this more so.

 * Standard C, POSIX, and system API support.
 * Basic text processing, a container package, logging facilities, and more.

Huzzah; I get to throw away my shoddy container implementations :)

 I think you will find few surprises beyond this, but I can mention
 non-blocking IO (selectors), ftp, http, locale, time and calendar support,
 some crypto algorithms. The rest is in the details, I think, and what we're
 still undecided on. There is a bit of useful template functionality in
 there too.

Well, you guys *have* been busy, haven't you? :3  I guess I'll just have 
to be patient (nuts).

	-- Daniel