The LOLCODE forum

Sigh, this apparently happened while I was at work.  (Getting my cluster up and running! W00t!)

Let's see...

Arachnid wrote:

"HAI" is now defined as "HAI [(TO|2) <version>]". Version is alphanumeric. The 1.0 reccommendation is "1.0" and this unofficial spec is "IRCSPECZ". If a compiler encounters a "HAI" with no version, it should attempt to compile it using the most recent/complete spec it knows. If it encounters a version it is familiar with, it should attempt to compile it as that version. If it encounters a version it is not familiar with, it should return an error and exit.

+1

"IZ" now permits any number of "MEBBE [IZ] <condition>" blocks, each introducing an 'else if' block. These come after the 'YARLY' block and before the 'NOWAI' block.

+1 MEBBE
-1 YARLY

The "YR" and "UR" keywords are treated as synonyms (they're used interchangeably by many people).

+1

GTFO is deprecated. (It was never actually in 1.0, but it's commonly used as if it is)
ENUF is introduced, serving the same purpose as GTFO. It can optionally take the name of the loop to exit: "ENUF [[OV [YR|UR]] <loopname> ]".

+1

MOAR is introduced, serving as a 'continue' statement. It optionally takes the name of the loop to exit, with the same syntax as ENUF.

+10! (+1)

A switch statement is introduced. This takes the form:

Code:

WTF [IZ] <condition>?
{ OMG <literal> [!]
    <statements>}
[ OMGWTF [?]
    <statements> ]
KTHX

Execution implicitly falls through from one "OMG" to the next until it hits an "ENUF", at which point it exits the switch statement. However, if a non-empty "OMG" block has no "ENUF" statement at the end, the compiler/interpreter should emit a warning to that effect.

+1

BTW is 'officially' the comment keyword (in this unofficial spec).

+1

"KTHX" to terminate loops is deprecated in favor of the old (and more flavorsome) "IM OUTTA YR <loop>".

+1

"DIAF" is deprecated in favor of "BYES".

-1 Code clarity, LOL-ishness!  I see these being used mostly without arguments, so BYES/DIAF are a lot clearer.  (i.e. "If file hasn't changed, BYES... if the file doesn't exist, DIAF!")

---

Comments are welcome. Once again, this isn't supposed to serve as an official spec, but as something unofficial with the less controversial extensions so that we implementors have something to be getting on with.

Signed,

Arachnid,
[other names omitted to protect the guilty. Take credit if you wish!]

Signed,
Eko

Last edited by etherealflaim (2007-06-07 02:30:36)

Arachnid wrote:

These are a series of extensions to the 1.0 reccommendation. They're entirely backward compatible, and entirely unofficial. Nobody is required to implement these by any stretch of the implementation, but we got tired of waiting. We tried to take only the least controversial proposed extensions and integrate them with the existing language. We naturally hope that this is what (part of) 1.1 will look like, but if it doesn't, we'll update accordingly.

+1

"HAI" is now defined as "HAI [(TO|2) <version>]". Version is alphanumeric. The 1.0 reccommendation is "1.0" and this unofficial spec is "IRCSPECZ". If a compiler encounters a "HAI" with no version, it should attempt to compile it using the most recent/complete spec it knows. If it encounters a version it is familiar with, it should attempt to compile it as that version. If it encounters a version it is not familiar with, it should return an error and exit.

+1

"IZ" now permits any number of "MEBBE [IZ] <condition>" blocks, each introducing an 'else if' block. These come after the 'YARLY' block and before the 'NOWAI' block.

+1

The "YR" and "UR" keywords are treated as synonyms (they're used interchangeably by many people).

Indifferent

GTFO is deprecated. (It was never actually in 1.0, but it's commonly used as if it is)

+1 Didn't like the language (trying to get lil kids here to laugh!)

ENUF is introduced, serving the same purpose as GTFO. It can optionally take the name of the loop to exit: "ENUF [[OV [YR|UR]] <loopname> ]".

+1

MOAR is introduced, serving as a 'continue' statement. It optionally takes the name of the loop to exit, with the same syntax as ENUF.

-10! I liked the idea of using "MOAR <variable>[!!#]" as a alternative to UP

A switch statement is introduced. This takes the form:

Code:

WTF [IZ] <condition>?
{ OMG <literal> [!]
    <statements>}
[ OMGWTF [?]
    <statements> ]
KTHX

Execution implicitly falls through from one "OMG" to the next until it hits an "ENUF", at which point it exits the switch statement. However, if a non-empty "OMG" block has no "ENUF" statement at the end, the compiler/interpreter should emit a warning to that effect.

+1, -1 I'm confused, if KTHX is depreciated, why use it in the example? Otherwise +1

BTW is 'officially' the comment keyword (in this unofficial spec).

+1 Comments are always good, though I'd like to see a quicker way for comments for us lazy typers Also, multi-line comments?

"KTHX" to terminate loops is deprecated in favor of the old (and more flavorsome) "IM OUTTA YR <loop>".

+1, -1 More stuff to type...

"DIAF" is deprecated in favor of "BYES".

+1 Never understood what DIAF was anyways

---

Comments are welcome. Once again, this isn't supposed to serve as an official spec, but as something unofficial with the less controversial extensions so that we implementors have something to be getting on with.

Enough comments?

Oh, if I can sign it, I do

Signed,
Zenix

zenix wrote:

I'm confused, if KTHX is depreciated, why use it in the example?

It's deprecated for loops, but it doesn't make sense to use "IM OUTTA YR <label>" for switch statements or IFs. Until we can agree on something more flavorsome, KTHX is it.

The irc channel is #lolcode on irc.freenode.net.

Church wrote:

Church's Bathroom Spec.

We the undersigned, being of minimal mind and/or patience, have hereby decided amongst my^H^H ourselves upon some extensions to 1.0 so that we may bigfoot our implementations.

Herein the proposals:

GTFO is deprecated. (It was never actually in 1.0, but it's commonly used as if it is. And we can't have that.)

Oh, don't worry your heads about it. I've got it all under control.

---

The Bathroom channel is #backroom on second.floor.can't.miss.it

If there's something you don't like about this, feel free to say instead of giving sarcastic commentary. Given that there are a bunch of implementors out there with little to do, and a bunch of proposals that most are in agreement over, and given that we're prepared to change what we implement when 1.1 is decided upon, should we have simply sat around doing nothing? Or all picked different features with slightly different syntax to work on while we wait?

-1 to MOAR as continue. Continue blatantly needs to be PROCEED

Heliomance wrote:

-1 to MOAR as continue. Continue blatantly needs to be PROCEED

+1

Last edited by Leo Natan (2007-06-09 21:22:57)

randomity wrote:

Why on earth must execution fall through from OMG to OMG? I can think of no sane reason that this should be so, or that anyone might want this, other than "it was in C". (BTW, the only reason it was in C is that it can be compiled down to some damn fast assembly language, not for elegance-of-language reasons)

This is going to bite n00bs. +1 everything else though, and +1 for this if OMGs stop execution at the next OMG.

Because without fall-through, a switch statement is just a glorified series of if-else statements. There are many situations in which you want to have several different cases execute the same code, and without fall-through on a switch, the only way to do that is by repeating yourself or with a huge multi-condition if statement. There's a reason every language that has switch allows fall-through.

This has also been hashed out thoroughly on the thread about the WTF statement.

Actually, now that I re-read it, the whole idea that you are deprecating things for flavor reasons indicates that this is not just a "stopgap" measure, but a full-on development of the language.  DIAF and BYES do the same thing.  Why not just leave it at DIAF until an official decision is made?  Or, if you want it to be BYES, why not create a proposal and make it so officially?

Peter

-1 MOAR for continue.
+1 PROCEED for continue.
+1 MOAR for incrementation.
-1 unofficial spec. It is not hard* to change the keywords around after building an implementation. What's the big deal?

* Edit: Actually it's trivial.

Last edited by superjer (2007-06-12 21:01:14)

I guess that's fine then. But you guys didn't go with what is currently the most popular and vote-winning. To the rest of us it seems fishy.

Forgive me, I was cranky.
It happens to the best of us.

I have a quick comment that probably won't get read about this whole forum proposal process and what you Forum only kiddies keep telling us implementers.  Actually, a few comments.

1) Without us, you have nothing.  Keep us bored long enough, and we'll leave and play with some other fun toy.  This is not a threat or me feeling self-important, it's just the way life is.
2) When we post things as proposals (like you CONSTANTLY keep telling us to do when you don't like what we've done), if they aren't contentious enough, people ignore them and they never get their non-implementer signoff (or implementer signoff, because they fall to the bottom of the index) and thus never get to a vote (and by induction, never make it to rec/spec).  When things get decided on IRC, they have plenty of implementer approval, but no approval from the forum people, so when they get here they get shot down and if a vote is proposed, you say it didn't satisfy the requirements and ignore it anyway.
3) The implementers know what they're talking about.  People don't sign on the channel and make decisions or votes that go into or really influence our irc-generated recommendations.  There is no accountability here on the forum, there is no motivation to have read the topics before you vote, and there is no incentive to peruse all of the previous threads to avoid rehashing things that have already been discussed.  And on forums, people HATE having a long post ignored or cast aside, because they spent a long time typing it.  On IRC, you get told really fast if you missed the big discussion and they've moved on, and/or that your idea has been accepted/rejected/proposed already and here is the biggest reason why.  Multi-player notepad is better, in my opinion, than a wall full of post-it notes.

[edit] It's late.  Spelling mistakes.  There are probably more.

Last edited by etherealflaim (2007-06-14 05:11:13)