home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Collection of Internet
/
Collection of Internet.iso
/
infosrvr
/
dev
/
www_talk.930
/
000143_davis@willow.tc.cornell.edu _Wed Jun 24 17:53:39 1992.msg
< prev
next >
Wrap
Internet Message Format
|
1994-01-24
|
2KB
Return-Path: <davis@willow.tc.cornell.edu>
Received: from dxmint.cern.ch by nxoc01.cern.ch (NeXT-1.0 (From Sendmail 5.52)/NeXT-2.0)
id AA28667; Wed, 24 Jun 92 17:53:39 MET DST
Received: by dxmint.cern.ch (dxcern) (5.57/3.14)
id AA03527; Wed, 24 Jun 92 17:54:04 +0200
Received: by willow.tc.cornell.edu (4.1/SMI-4.1)
id AA13109; Wed, 24 Jun 92 11:55:33 EDT
Date: Wed, 24 Jun 92 11:55:33 EDT
From: davis@willow.tc.cornell.edu (Jim Davis)
Message-Id: <9206241555.AA13109@willow.tc.cornell.edu>
To: www-talk@nxoc01.cern.ch
Subject: Re: Links, Types and Documents
In considering R. Raisch's link types, it seems to me that the
expansion and replacement can be combined and generalized. In both
cases portion of the document is replaced. In expansion case, the
condensed text is replaced (perhaps temporarily) with the enlarged
text, in replacement the entire document is replaced with some other.
So the more general case is a link which labels a range in both the
source and target document. The replacement link labels the entire
document, the expansion labels only a portion.
As for execution links in general, I see Dan's point that adding
procedures to documents makes them difficult to manipulate. And N
Borenstein has pointed out (in Mime) the dangers of executing
arbitrary code. But I still see that there are functions which only
code can provide, and so we just face the challenge of designing a
language that is both powerful and safe, perhaps sacrificing Turing
completeness. Is Atomicmail [tm?] such a language?