Re: Question about Sun JAVAC
From: Edward G. Nilges (spinoza1111_at_yahoo.com)
Date: 09/09/04
- Next message: Edward G. Nilges: "Re: [EGN] Re: Hackers"
- Previous message: Edward G. Nilges: "Re: Hackers"
- In reply to: Michael Wojcik: "Re: Question about Sun JAVAC"
- Next in thread: Michael Wojcik: "Re: Question about Sun JAVAC"
- Reply: Michael Wojcik: "Re: Question about Sun JAVAC"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 8 Sep 2004 22:14:30 -0700
mwojcik@newsguy.com (Michael Wojcik) wrote in message news:<chn74g01qh9@news2.newsguy.com>...
Thanks for an interesting post. However, regular expressions remain a
language adopted eagerly in 1970 owing to their expressiveness
relative to hardware and today they gnomically perform functions that
could be performed by more expressive notations, notably BNF.
> In article <f5dda427.0409061533.46d0765@posting.google.com>, spinoza1111@yahoo.com (Edward G. Nilges) writes:
> > mwojcik@newsguy.com (Michael Wojcik) wrote in message news:<chi6i1027as@news1.newsguy.com>...
> > > In article <f5dda427.0409021555.748255dd@posting.google.com>, spinoza1111@yahoo.com (Edward G. Nilges) writes:
> > > >
> > > > This is because regular expressions are a form of machine language
> > > > that was designed around the limitations of input and output devices
> > > > circa 1970.
> > >
> > > Regular expressions are a mathematical construct that date back at
> > > least to the work of Turing and Kleene. That was significantly prior
> > > to 1970.
> >
> > Yes, but the current encoding is based on the limitations of input and
> > output devices circa 1970.
>
> The RE and ERE syntax used by various Unix utilities clearly was
> developed using the I/O devices common to Unix systems of that era,
> and operating within their limitations was certainly a requirement;
> but it employs a syntax not significantly different from the
> preexisting mathematical one (or ones, since there was some
> variation, for example in using "+" or "|" for alternation).
>
> Of course the symbols available for ASCII I/O devices (that is, ASCII
> itself, sometimes limited by unfortunately under-featured terminals)
> represented the ANSI and ISO committees' best efforts to arrive at
> the most useful compromise among the sometimes-conflicting needs of
> various potential users. Thus ISO initially made the tilde and hash
> signs alternates for the same ASCII code point - a situation which
> would have changed the syntax of C, for example, had it remained.
>
> In other cases, one symbol served multiple masters. Jukka Korpela
> notes that Bob Bemer told him that the asterisk was initially
> introduced into many pre- standard computer character sets for
> commercial purposes, such as "check protecting" (filling in white-
> space around numbers on checks to guard against alteration). Only
> later was it adopted by programming languages for multiplication and
> other mathematical functions.
>
> And some symbols were added for one (sometimes relatively obscure)
> purpose and later adapted to others. The classic ASCII example is
> the backslash, which Bemer initially advocated for use in concert with
> the slash to represent Boolean conjunction and disjunction (as "\/"
> and "/\") - there's a well-pedigreed ASCII syntax for your symbolic
> logic. It was quickly seized on to represent "reverse division" as
> well, though there apparently wasn't a consensus on just what
> "reverse division" might entail. (Apparently some languages used
> it to represent continued fractions: a\b\c...) Then, of course,
> Microsoft (in)famously used it as a path separator in MS-DOS 2.0,
> because the slash was already taken as option separator.
>
> But traditional mathematical syntax for regular expressions happened
> to use characters that were already in ASCII - not entirely a
> coincidence, of course, since the people who developed RE syntax
> elected to use already-established symbols rather than introducing
> novel ones, and ASCII incorporated symbols that were for the most
> part already widely used. The most notable feature of RE syntax, the
> Kleene closure operator, has always (AFAICT) been written with an
> asterisk (which is why it's often called the "Kleene star", or by
> wags the "Kleene X", though that pun depends on mispronouncing
> Kleene's name). Thus we cannot in fairness claim that ASCII I/O
> devices constrained Unix RE syntax; rather, they facilitated adopting
> the syntax already in use.
- Next message: Edward G. Nilges: "Re: [EGN] Re: Hackers"
- Previous message: Edward G. Nilges: "Re: Hackers"
- In reply to: Michael Wojcik: "Re: Question about Sun JAVAC"
- Next in thread: Michael Wojcik: "Re: Question about Sun JAVAC"
- Reply: Michael Wojcik: "Re: Question about Sun JAVAC"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|