Log of the #nice channel on calvino.freenode.net

* grom joins05:43
did anyone here see my initial ideas for properties??05:45
* grom leaves06:37
* arjanb joins09:22
* Quick-Nic leaves10:03
* GummiEnte joins10:15
Hi@all
Seems so, as if nice community starts to grow.10:16
What's up with the logger... It's sometimes a few days offline?!
<arjanb>hi11:07
the computer where the logger runs on is shutdown sometimes11:10
and do you have a problem again?11:11
<GummiEnte>No, I'm pleased with using nice.12:28
<arjanb>great, much code have you written in Nice?12:29
<GummiEnte>Well, currently I assume about 4000 lines... (with comments, so you can only speak of quite the half of real lines)
Well, up to now, I've encountered no great performance issues, but I only use the classes to the let gcj compile them nativly.12:32
I've read in the logs, that no one really has written a large program with nice?12:37
<arjanb>not that i know of12:38
<GummiEnte>Why not?
I think I'm going to use it instead of java in the future.12:39
<arjanb>i have no idea why no larger program's are written yet12:40
maybe we just never hear about it or people stop after trying out nice12:42
<GummiEnte>Well, the first try may be hard... Coming from java it is first an uncommon think to compile only packages and not classes.12:44
<arjanb>yes isaac pointed out that too i hope the documentation is more clear about that now12:51
* bonniot joins13:07
<GummiEnte>I've read something about bootstrap compiler's in the log...
<bonniot>hi
<GummiEnte>Hi
What about that?
<bonniot>i'm just planning a modification that will make bootstrap work more smoothly when a new inline operation is added to the compiler13:09
regarding Nice usage: you have to keep in mind that 0.8 was released only three month ago, and that almost no announcing was done before that13:11
we can see the growth starting really now
but adoption of a new language always takes time, that's normal13:12
there are lots of aspects that can still be improved to make adoption easier, they just new workforce :-)13:13
<GummiEnte>Well, get me right. I'm completly satisfied with nice. I'm not one of the ons, that extract the software, klick on it and expect it to run, that I wanted it too. There might be some other stumbling blocks.
One should carefully read the documentaion...13:14
<bonniot>you think the installation is hard?13:17
<GummiEnte>No, completly not. Just really easy. But the first steps are for guys, that don't read a bit of docs a real hurdle.13:18
...the first steps in using the compiler and writing nice code...13:19
<bonniot>yes, this was discussed on nice-info13:22
if you have suggestions...13:23
surely if they don't read any doc it will be hard
<GummiEnte>:) what should you give for suggestion instead of reading docs...
<bonniot>but the docs could also surely be improved
<GummiEnte>Ok, that might be...
<bonniot>i just don't have so much time13:24
<GummiEnte>The Wiki is good for developers and suggestions but for me personally it is not most intuitive way, because you might need to look into all hooks of the wiki to gain what you are searching for...13:26
<bonniot>if it is well organized, you should be able to follow a tree structure to find information13:28
you can also use the search facility
<GummiEnte>But if I go to the start, I have to decide which group to use. Quite a few infos are only contained in devlopment group but are also interestng for me. If you are interested in nice you will find what you want, but maybe it is not the fastest way and so someone may drop the search...13:31
It fits my needs, but what with the guys, that are so lazy to read and search... They better go to write a mail to the mailinglist and ask ...13:32
<bonniot>Normally it should be Doc for the users
You are an advanced user, if you are interested in Dev ;-)13:33
<GummiEnte>:)
<bonniot>What did you find there?
<GummiEnte>Well, Doc is sufficient.13:34
http://nice.sourceforge.net/cgi-bin/twiki/view/Doc/GetInvolved <-- This site I've never seen... 13:35
NiceUnit <-- what is meant be that?13:37
The nicedoc would also be interesting.13:42
<bonniot>Yes, GetInvolved should be mentioned on nice-info13:44
NiceUnit is a unit tester. Haven't you heard of jUnit?
<GummiEnte>Well, to be honest, no. But I've hold a seminar about ExtremeProgramming.13:45
<bonniot>so it seems you know what it's about ;-)13:48
<GummiEnte>Google helped me a bit to reference Junit to xp.13:49
<bonniot>:-)13:50
<arjanb>hi daniel13:58
<bonniot>hello arjan14:02
<arjanb>any progress finding the coverage bug?14:03
<bonniot>no, i'm working on my thesis at the moment
and you?14:04
<arjanb>thinking what to do next
<bonniot>whyt about visibility?14:06
<arjanb>not now it affect's to much source at once14:09
<GummiEnte>:) I now what arjan means. Better let the code as it is, because it is working. But once it as to be done...14:10
<arjanb>it possible first step is splitting the global scope into global package and file scope14:11
that will also make correct sematics for the import definitions possible14:13
<bonniot>what I would like is to see a symbol visibility package developed independantly of the compiler14:31
I can do the job of modifying the compiler to use it if you wish
<GummiEnte>so, with an independent package, it would be possible to use the compiler with and without visibility?!14:32
<bonniot>no, it's about modular development14:37
i don't think it would be useful to disable visibility
<GummiEnte>Ok.14:42
<arjanb>i don't see anything in visibility independantly of the compiler14:50
i have looked at other compilers such as the pizza compiler and it makes me think that visibility things will end up in a lot of places14:52
<bonniot>it depends what you mean by "things" :-)14:57
there are calls to the API, which will be in bossa.syntax14:58
and the implementation, which can be made in an independant package
* roerbak leaves14:59
<bonniot>imagine that we had two different compilers. we could want them to use the same symbol handling library
so that has to be in a separate package
<arjanb>so you are saying to move scoping and overloading together with visibility in a seperate package15:04
<bonniot>scoping and visibility
lookup of a symbol can return several results when it is overloaded15:05
the resolution has to be in the code, since it needs typechecking
the resolution of overloading
<arjanb>so your goal is to split up the compiler as much as possible to make it easy extensable?15:07
<bonniot>yes15:08
and also easier to understand and modify the code, and to test the different parts individually
<arjanb>i see for me looks seperating the parts closely related to the conversion to Nice source code15:11
is replacing java as target of programming languages research extensions a long term goal?15:12
<bonniot>1) the scoping package can be written in Nice, so it's part of the conversion15:13
2) it's not a main goal for me, but making the compiler more modular is good anyway, and yes it could encourage that15:15
* arjan_b joins16:03
* arjanb leaves16:22
daniel: "I plan to remove functions and the syntax for method implementation." which syntax do you mean?17:54
<bonniot>m(A x) instead of m(x@A), and the possibility to override the return type:18:01
override C m(C x) { ... }
<arjanb>i have doubts about removing patterns like x@A18:04
<bonniot>can you be more precise?18:07
<arjanb>using @ and # makes it clear that you want dispatching on the type18:09
<bonniot>what is not clear about (A a) ?18:11
<arjanb>the difference between an implementation and declaration becomes less
<bonniot>but there is still a difference18:12
the difference is mainly that a declaration has no code
that's very obvious18:13
then there is the declaration with a default implementation
<arjanb>and then the only difference is having a return type or not18:14
<bonniot>yes18:15
<arjanb>but for readability you may want a return type for implementation too
<bonniot>then you must use the keyword override, which makes the difference obvious
<arjanb>but you don't override the return type18:17
<bonniot>possibly you do, but in any case you override the method18:18
<arjanb>i think i going to write down my ideas about this so we could discuss this with others more easily18:23
can override mean contravariant argument types too?
<bonniot>ok18:34
no, i don't think contravariant arguments would be very useful, and it would be hard to find out what method is overriden18:35
<arjanb>yes18:40
hmm virus attachment in alex's wiki notification mail20:13
<GummiEnte>There are quite a lot viruses and spam around the world the last weeks20:14
So, I'll now...20:27
* GummiEnte leaves20:30
<bonniot>strange: I got the notification mail for Doc22:56
Then Potentially dangerous email rejected22:57
oh, it's a message from your uni in .nl
<arjanb>yes
<bonniot>i don't see anything wrong in the notification email22:58
either it's a bug in the checker, or the message was modified on the way
i think
<arjanb>the checker is a bit strict i think23:00
At Fri Sep 5 20:00:18 2003 the content filters said:
Found a form in HTML message
<bonniot>is it a new filter, or the message format has changed?23:09
i suppose the former23:10
you should complain to your admin ;-)
<arjanb>the message format hasn't changed
<bonniot>in your change to NullExp, you left some commented code23:28
<arjanb>forgot to remove that23:29
* bonniot leaves00:11
* arjanb leaves00:24
* bonniot joins00:30

Generated by Sualtam