[Oberon] Oberon-3

[Oleg N. Cher]

Isn't it defined in Component Pascal whether ORD ({0}) = 1, and not 31, 
or 63 ? (I do not know how this happens in Oberon-07)


Jörg пишет:
> Oleg
> 
> There is no doubt ORG(set) is useful; my color Display driver is full of them.
> But the strange thing is that NW allows it, although it is HW and implementation dependent.
> 
> It's not defined whether ORD ({0}) = 1 or 31 or 63. And it is not defined whether {0} is the leftmost or rightmost pixel.
> And you can do all this without importing SYSTEM...
> 
> br
> Jörg
> 
> Am 04.11.20, 13:08 schrieb "Oberon im Auftrag von Oleg N. Cher" <oberon-bounces at lists.inf.ethz.ch im Auftrag von allot at bk.ru>:
> 
>     Hi Jörg,
> 
>     Unfortunately, Component Pascal does not support ORD() for BOOLEAN. 
>     Since we have the BITS() operation there, it is good to have a pair 
>     operation ORD(set) for converting bits to a number.
> 
>     In Oberon-3 I supported ORD() for SET and BOOLEAN, as found it useful.
> 
>     Ilya Ermakov noticed that ORD(procedure) is useful as a unique stamp 
>     number for unambiguous identification of procedures. Such operation does 
>     not look dangerous to be moved into SYSTEM, although it can hardly be 
>     called portable. Besides this can cause theoretical problems if ORD() 
>     returns a 32-bit result and a system has 64-bit address space.
> 
>     So in general, I wasn't going to implement ORD() for other types than 
>     these. Although, of course, it will be interesting to hear different 
>     opinions on this issue.
> 
> 
>     Jörg пишет:
>     > Oleg
>     > 
>     > I read in your Oberon-3 description that you support ORD() on several types.
>     > Oberon-07 defines ORD() on CHAR, BOOLEAN and SET.
>     > NW's compiler supports ORD() on more types as well, but this is undocumented.
>     > Generally, ORD is a little dangerous, as ORD returns implementation dependent values, and hence a code using ORD might not be portable.
>     > 
>     > NW's Oberon-07 compiler (ORP.StandFunc) supports ORD() on
>     > BYTE, BOOLEAN, CHAR, INTEGER, REAL, SET, POINTER, NIL, PROCEDURE
>     > 
>     > Puristically, it is better to use SYSTEM.VAL(INTEGER, x) instead of ORD(x) as SYSTEM flags the implementation dependency.
>     > 
>     > It is again (like the semantics of suffix "H") one of the undocumented compiler features NW uses in his own code.
>     > With a strict compiler implementing ORD() according to the Oberon-07 report, you cannot compile Texts.Mod and System.Mod in its current form.
>     > 
>     > br
>     > Jörg
>     > 
>     > Am 04.11.20, 10:47 schrieb "Oberon im Auftrag von Oleg N. Cher" <oberon-bounces at lists.inf.ethz.ch im Auftrag von allot at bk.ru>:
>     > 
>     >     August Karlstrom пишет:
>     > 
>     >     >>    Oberon-3 (experimental dialect with constant arrays and "proper FOR")
>     >     > 
>     >     > Oberon-3 seems to be so obscure not even Google can find it.
>     > 
>     >     I may have shown some impudence in taking this name. But I didn't plan 
>     >     to make my own Oberon dialect, I just needed some extensions for my 
>     >     activities. I hope this doesn't cause too much anger among those who 
>     >     respect Prof. Niklaus Wirth and his associates.
>     > 
>     >     In any case, this development is unlikely to spread too widely and will 
>     >     become very popular. So the name "Oberon-3" can be considered vacant for 
>     >     further achievements.
>     > 
>     >     Thanks for understanding.
>     > 
>     > 
>     >     > 
>     >     > 
>     >     > -- August
>     >     > -- 
>     >     > Oberon at lists.inf.ethz.ch mailing list for ETH Oberon and related systems
>     >     > https://lists.inf.ethz.ch/mailman/listinfo/oberon
>     > 
>     >     --
>     >        Oleg N. Cher
>     >     --
>     >     Oberon at lists.inf.ethz.ch mailing list for ETH Oberon and related systems
>     >     https://lists.inf.ethz.ch/mailman/listinfo/oberon
>     > 
>     > 
>     > --
>     > Oberon at lists.inf.ethz.ch mailing list for ETH Oberon and related systems
>     > https://lists.inf.ethz.ch/mailman/listinfo/oberon
> 
>     --
>     Oberon at lists.inf.ethz.ch mailing list for ETH Oberon and related systems
>     https://lists.inf.ethz.ch/mailman/listinfo/oberon
> 
> 
>