Quote:

Wrong. The PowerPC manual uses 'word' to signify a 32 bit machine word, NOT a 16 bit word.

Sigh... and where did I mention the PPC? There was quite a big reasoning there, fully based upon the fact that the 68k views the types that way. The PPC was completely out of the picture.

Perhaps you should stay more calm, when replying to me.

Quote:

So basically you are upset because we didn't copy AROS' names?

Again, I'm not upset, the upset one is you here.

Quote:

Sorry, but what if we had named it QUAD?

Quite simply, I wouldn't have had anything to complain about, of course.

Quote:

You would probably go on to complain we copied from AROS, wouldn't you (don't answer, it's a rhetorical question)

Err... so you are free to say absurd things about me, and I'm not even entitled to tell you "Sorry, but you're totally wrong"? Ok, fine, if you like to believe the world is out to catch you, it's your freedom to do so.

But then don't expect people to take you seriously.

Quote:

You're upset because you where "there first" and we didn't "respect" your choice?

I did say "there first", but I did not even mention the word "respect", so I would be thankful to you if you made it clear that's not an exact quote, but just your (faulty) interpretation, thanks.

Quote:

Sorry to say so, and you're probably going to call me arrogant now (you wouldn't be the first for today ), but AROS' decisions are AROS' decisions, and not AmigaOS 4's decisions. You might think that QUAD is a good idea; we don't. Like I said (repeatedly now) a word in PowerPC means 32 bits. It was therefore the only viable alternative to get rid of this misleading thing.

"It was therefore the only viable alternative" is your subjective view. Obviously, AROS didn't deem it the only viable alternative, nor did Windows.

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

Quote:

WORD does not mean 16 bits in AmigaOS 4; the typedef is reproduced for compatibility reasons, and will vanish as soon as possible.

Wait wait wait... didn't you just say that WORD is there to stay, for sake of 3.1 compatibility? Taking it back?

Quote:

Plainly, we make the standards.

Quote:

We listen to input as much as we can, but some decisions are simply made. I am not going to tell you that QUAD in AROS is a problem, but in AmigaOS a 64 bit quantity is an int64 (signed) or uint64 (unsigned).

But you should tell me QUAD in AROS is a problem: if QUAD is a problem in AOS4, it must be a problem in AROS too, must it not? However, QUAD is not a problem in AROS, it has never been in the past 7 years, it's not because you think it's a problem that suddendly it becomes a problem.

You're free to chose whatever you want to use, just like I'm free to question whatever I think I want to question. This is a free world, after all, don't get upset so easily

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

Quote:

Someone was there first in 1985 and found the 68k CPU as a solution.

Certainly not a solution to the problem we're talking about here, so I don't see why you mention it.

Quote:

Where does that get us? Nowhere. The 68k is quite obviously dead as a doorknob. Time to move on to a different solution, obviously. I don't understand why you need to make such a fuss about it;

You fail to understand that what CPU is underneath the OS is irrelevant. This is about source compatibility, and established semantics. Changing them for no gain is useless. If it ain't broken, don't fix it.

It really sounds like your doing this change for the sake of it.

Quote:

obviously AROS has taken a different route in the first place, so what are you worrying about? Or is this just about your Ego, because you think we should have followed the same decision as you did?

I think you should have followed an established route, and not take a different one just to be different and tell the world you make standards. Get what I mean? More often than not you seem to reinvent the wheel for no good reason.

AROS in itself has nothing to do with it, as said you could have chosen to use the standard c99 types and I would have had nothing to complain about, but you are changing one non standard nomenclature (but standard within the AmigaOS set of apps) for another non standard nomenclature (not even within the AmigaOS set of apps): what's the gain? I can't see any.

Quote:

Now you are becoming ridiculous.

Thanks a lot. Really.

Quote:

It isn't as if WORD and LONG and friends are "standard" by any means.

See above.

Quote:

They're used exclusively in AmigaOS.

And the new ones will too.

Quote:

The old typedefs are still there, so you can still use them for legacy code, but other than that, the new naming scheme *is* the new standard for AmigaOS.

Changing your mind again? First you say they will always be there, then you said they won't be there anymore, and now you say they'll be there again. I'm confused.

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

Quote:

Sigh... and where did I mention the PPC? There was quite a big reasoning there, fully based upon the fact that the 68k views the types that way. The PPC was completely out of the picture.

I'm sorry, but it was you that brought up this issue in the first place. You asked for the reason to deprecate the old types, you got an answer, and then started to argue about it. The 68k is of no concern to AmigaOS 4.0, and that is what this thread is about. It was you that started arguing that this was a bad idea, and that WORDs where 16 bits.

So, pray, why is the PPC out of the picture? AmigaOS 4.0 is all about the PPC.

Quote:

Obviously, AROS didn't deem it the only viable alternative, nor did Windows.

Like I said, both AROS and Windows are completely irrelevant to this thread.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

Quote:

Wait wait wait... didn't you just say that WORD is there to stay, for sake of 3.1 compatibility? Taking it back?

Look up the definition of "deprecated".

Quote:

But you should tell me QUAD in AROS is a problem: if QUAD is a problem in AOS4, it must be a problem in AROS too, must it not?

Plainly I don't care about AROS. I said that we made the choice to deprecate the old types, period. As I said before it was you that started to argue about it being a bad choice.

Quote:

However, QUAD is not a problem in AROS, it has never been in the past 7 years, it's not because you think it's a problem that suddendly it becomes a problem.

Beg your pardon, where did I say something like that? I guess that I didn't mention yet that I don't care about AROS, and in fact I was always talking about AmigaOS in the first place. Again, I don't care what you do and how you do it in AROS, and I most certainly don't try to tell you how you should do it.

Quote:

This is a free world, after all, don't get upset so easily

What makes you think I'm upset? I think you are taking yourself too serious. I am just trying to reason, and try to contradict where I think you are wrong. Which happens to be in quite a few places.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

Hi

I'm not an advocate of OS4 developers, but this certainly has a clue.
The [u]int[8|16|32|64] is certainly better than DWORD, QUAD, naming because:

1. It creates much cleaner _mechanism_ for naming of types than: DWORD, QUAD, etc.. and creates more flexible mechanism for the future (although, i hope not too much new atomic types are introduced ).
2.Cleans up the mess with system types and with compiler types. As I see the set of types of the compiler is now a subset of the system types. So the developer can (almost ) see one set of types. Correct me, if I'm wrong.
3. easier to write, no shift usage
4. I'm an advocate of lowercase naming of not object types, so this complies

Good job OS4 devs. Should be done sooner.
Microsoft did the same in .NET Framework (but that's not why it is good )

re

Treke

Quote:

Certainly not a solution to the problem we're talking about here, so I don't see why you mention it.

I mentioned it because it serves as an example that whatever the solution was yesterday it doesn't mean that this holds true for eternity.

Quote:

I think you should have followed an established route, and not take a different one just to be different and tell the world you make standards. Get what I mean? More often than not you seem to reinvent the wheel for no good reason.

"Established"? QUAD is not an established standard just because it has been used in AROS.

I have given the reasons for not using something like QUAD, and I personally think that they are good reasons. You may think different, but there is a difference between changing something and "reinventing the wheel".

Once again, for clarity:

- I think that using names like WORD is misleading because every CPU has its own idea what a WORD is. AmigaOS may have a 68k legacy, but that is exactly that - legacy. Newcomers might not grasp this.

- On the PowerPC (which is the CPU that runs AmigaOS 4.0) a word signifies 32 bits.

- The type names used now are completely unambigous.

- Using C99 types was considered, but my opinion is that there should be a difference between system header types and "standard" types. For starters, the AmigaOS types are language-neutral. We musn't look at C/C++ only. IMO it seems natural to make a distinction between system types and language types.

Quote:

And the new ones will too.

Yes. This is intentional.

Quote:

Changing your mind again? First you say they will always be there, then you said they won't be there anymore, and now you say they'll be there again. I'm confused.

Sigh. I said they are still there, and they will stay for a time, but "deprecated" means they are discouraged from further use and may vanish at any time. You can, however, still put them into a compatibility header. I never said "they will always be there", I said they are "still there" as in "still there now".

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

Quote:

2.Cleans up the mess with system types and with compiler types. As I see the set of types of the compiler is now a subset of the system types. So the developer can (almost ) see one set of types. Correct me, if I'm wrong.

Not quite. C99 defines types like uint32_t, but IMO a system API should not use these as they are language-specific.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

Quote:

So, again, why didn't you use the standard c99 types? I'd have nothing against that, as long as you kept the standard AmigaOS types around.

Ok, no need to answer to that, I read your other replies too.

Quite frankly, it's not like I would give a damn about this whole issue, it all started with an innocent question. As it often happens, I don't agree with many decisions you guys make, it must just be a consequence of the fact that we may be looking at the world with a different set of spectacles.

Let's put an end to this, as I'm exhausted... only, please, in the future, avoid making bold claims about me, my motives and the rest of the world in general: it doesn't come out very good.

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

Quote:

Look up the definition of "deprecated".

I know what it means, perhaps it's you the one who's not got it clear. You explicitely stated, at first, that those types are there to stay, except you took it back later on.

Quote:

Plainly I don't care about AROS. I said that we made the choice to deprecate the old types, period. As I said before it was you that started to argue about it being a bad choice.

No one asked you to care about AROS, however if you think QUAD is a bad choice, then it is regardless of the OS, so it's a bad choice for AROS too, in your view.

However, 7 years should be an enough long time to show that QUAD is not a bad choice. Wouldn't you agree with that?

See, this is the proof that QUAD is not a bad choice. I don't know how else to explain it to you. You can of course say that you don't like it, but don't say it's a bad choice, don't dare say yours is the "only viable solution", as those are quite bold, and unsubstantiated statements.

Quote:

What makes you think I'm upset?

Hum... let's see... how about : the same thing that makes you think I am upset?

Quote:

I think you are taking yourself too serious. I am just trying to reason, and try to contradict where I think you are wrong. Which happens to be in quite a few places.

That's funny, as I happen to think the same about you! How's that?

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

Whoops, I edited the post instead of replying to it... Wrote all that for nothing :-/ Well, this will teach me to mind my business, next time, and "not care" about AmigaOS4 at all

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

@falemagn
Quote:

This is about source compatibility, and established semantics. Changing them for no gain is useless. If it ain't broken, don't fix it. It really sounds like your doing this change for the sake of it.

When you are migrating your code to OS4, you'll review the migration guide and have the choice of changing many things beyond just the basic OS types. If you choose to use the old names, your source code remains compatible and your established semantics haven't changed. No problem.

If you are talking about creating brand new OS4 code, I think you should use the new basic type names. They are unambiguous, hardware independent, programming language indepedent and have mostly unique labels.

For mixed old and new OS4 projects you'll have a more difficult time. Do you go through and change all your old types to the new system or keep using the old types? That's a decision each OS4 developer will need to make on their own.

I recommend trying the new type names in your own OS4 projects for a while before passing judgement. I've been using the new type names myself for a while and I don't have any problems with them.

_________________
ExecSG Team Lead

Quote:

True: 4 words. What's difficult about it?

Sheesh, stop being so obtuse!

A WORD on 68k and x86 is 16-bits. Therefore for you and AROS, it is fine to use that.

However, on PowerPC, the native word length is 32-bits.

So there you go. "4 words. What's so difficult about it?" means a 128-bit integer.

I think that Hyperion's definitions are a lot clearer and sensible, and show that they thought about the issue and the future.

Maybe they'll have QFloat32s, OFloat16s, etc, for Altivec, heh.

@falemagn:

Quote:

Sigh... why is it that it's almost always impossible to have with you a calm discussion about something? Did you really need to question my programming abilities, or even if I ever went to a "programming class"? Besides, what has that got anything to do with what we're talking about?

I guess you have a way of rubbing me the wrong way... sigh. I try not to be rude, but there is a small set of people who happens to bring out the worst in me, and unfortunately you seem to be part of that set. Please don't take anything I say too seriously, and if something I say could be interpreted in two ways, and one of them is an insult, I meant the other one

Although I admit I could have worded it better, I wasn't implying that you don't know how to program. Rather, I was implying that you've always known how to program (the boy genius remark), therefore never set foot in a classroom, therefore have no idea that there are people out there who do have problems with things like the same word in a different upper/lower case having different meanings.

Thanks for the snappshots!

Just to get back on topic, I like the idea of having an installer that sets everything up for a developer and it all looks great! Hyperion have done an excellent job of putting this all together.

P.S. I can hardly wait to see how much feed back there is once OS 4 pre-release hits our A1's. People rarely like change, even though it is for their own good. Just ask God, who gives us everything we need to be happy and yet we still want what will ruin us. As a side note, I like the new syntax and find it much easier to understand.

@falemagn,

At the end of the day new naming conventions to describe the different memory sizes have been introduced. In my opinion the new standards are for the better; they are clear, leave room for obvious expansion (as and when new sizes are required), and steer clear of names which have different meanings for different architectures. In your opinion they are not brilliant choices for your own reasons mentioned above ^ ^ .

The fact is; new standards are here. This is what OS4 programmers will be using from now on (unless they use the legacy WORD, LONG, ULONG etc..). Hyperion have layed down the new standards for us all to use when we start programming with AOS and they make sense.

So we may have different opinions on this matter but it's not designed to be incorrect or unusable.

_________________
OS4.1 + SAM Flex
RIP my A1XE.. that used to have an appetite for batteries!

@ nzv58l

Yep, looking forward to the dev CD. This is getting quite exciting really!

_________________
OS4.1 + SAM Flex
RIP my A1XE.. that used to have an appetite for batteries!

Back in the 80's I took computer courses on a VAX 11/780 system, and the system specified integer data type sizes as the following:

Byte 8 bits
Word 16 bits
Longword 32 bits
Quadword 64 bits
Octaword 128 bits

Personally, while Quadword was the first thing that came to my mind for the new AmigaOne/OS4.0, obviously the different interpretations of "word" between the 68K and PPC processors does make [u]int[8|16|32|64] the perfect choice, leaving no room for misinterpretations.

The whole int, short, long, word, ... issue has one problem: We have to think like machines.
Hyperions choice IMO is the best one. I even would have liked this naming convention for Java. What if 64 bits isn't enough anymore? What if in 10 years we have standards of 256 bits? How will we have to call those datatypes? What if somebody is devolping for Amiga, MacOS, Linux, Solaris, Windows? Do we always have to remember which name means how many bytes on which processor? I am a software developer, but I am still a human being. And I don't want to think like a machine.
Fabio, you stick too much to the machine. REMEMBER: Machines are there to help us, do our work. We don't want the machines to determine our way of thinking. This is not good.
Why use an OOPL? Because it lets us think in terms of problems. What is a program? What is an algorithm (even in mathematics)? It is a solution to a problem. So, why should we convert our thoughts into bit-streams? Let's stay in the problem space and you'll find a cleaner solution.
Everything is about being INTUITIVE. Your solution has nothing to do with intuition. It's only nice for you. Why? Because you are used to WORD = 16 bits. But being used to something is not the same like using something in an intuitive way.
We are both italians. Think about what makes our language so easy to learn? Take for example the pronunciation. It is straightforward. It's intuitive! You have some rules and with those you can pronounce every word correctly. You don't have to invent new rules. Even new words (in processors they could be 16, 32, 64, 128 bits) are always pronounced using the same rules (in programming [u]int[8|16|32|64|128), you see?

I know that you are a way better software developer than me. I am also some years younger than you. Nevertheless, I think that you should read carefully my comment and see if I am at least in some parts right.

Quote:

Whoops, I edited the post instead of replying to it... Wrote all that for nothing :-/ Well, this will teach me to mind my business, next time, and "not care" about AmigaOS4 at all

Hey, come on! Think of how you could contribute to a system you learned to love. No, don't say you don't love Amiga anymore. You even started a "port". Please don't be offended if this is not the correct definition of AROS. This is only done by people who love a certain system. AmigaOS4.0 is the new release. New releases take changes with them. Don't be offended by comments of other people. Think of how often you offend people next to you or in forums. See it as a discussion from which everybody has learned something.
Even you teached us something: Nobody, who read this whole thread will ever forget how many bits an int, a WORD, long, LONG, ... have.
Remember, developing for Amiga is a lot more fun than developing for Windowze!

My opinion is that several changes had to be made in AmigaOS3.x to bring it to 4.0 And Hyperion made a great job.
Maybe this should be put in a new thread, but I want to add it here:
@Hyperion: THANK YOU! Many people say this is not Amiga anymore. In this case we could also like WindowsXP. It's not Windows anymore, right? NO! WindowsXP is the Windows release that followed 2000. And it' still Windows! AmigaOS4.0 is still AMIGA! With hardware abstraction, some memory protection, refreshed GUI, .... All things we needed. Maybe something is not that useful? And some other tools are necessary? Hey, we have also 4.1, 4.2, ...
I think we should accept some changes and learn to do the best of it.