[lojban] Re: open and save

[Jim Carter <jimc@MATH.UCLA.EDU>]

On Tue, 7 Jan 2003, Steven Belknap wrote:
> On Monday, January 6, 2003, at 11:53  PM, Robin Lee Powell wrote:
> > <despair>
> > Nooooooo!
> > </despair>

> The use of the lojban word <dacru> for a computer file is not malglico.
> The analogy of computer file to a paper file is a language-independent
> extension of the concept of file to cyberspace. <vreji> is not an apt
> lojban word for file. A file *contains* records.

I think that we've blundered back into the "wine bottle / bottle of wine"
debate, just in a different area.  Container words are generally defined as
"X1 is a [container] of [content] X2..."  So pretty clearly, "se
[container]" refers to the content, provided it is actually in a container
of the specified kind.  But does X1 refer to the container with its
content, or the container alone?

Before the file is opened, the focus is on the container with its content.
After it is opened, you want to work on the content and not be bothered
with the container.  I find "dacru" unaesthetic as a metaphor for "file",
but I suppose it's useable, having the right arguments, though not in the
standard order (container - content - etc).

As for vreji "x1 is a record of data/facts x2 about subject x3 medium x4",
I think "ve vreji" is rather better as a symbol for the file as a container
with its contents, independent of the context, e.g. computers, paper, stone
tablets...  Then, after the file is open, "mi tcidu le vreji".

The "medium" can equally well be "paper" or "the thing that the ReiserFS
code looks for on your hard disc partition when you pass its name to the
"open" system call".  It's not necessary to be anally retentive and say
that a medium has to be first order physical, so that "hard disc" is the
only acceptable value.  Also remember that a file can contain actual files,
e.g. a "tar" archive, and some software can access the contained files
transparently.

I can understand Steven's point about "record": in many but far from all
files, the file contains an organized sequence of sub-files, called
"records" in English, each of which give a variety of data about one
subject x3.  I looked for good gismu for "stanza" or "segment" (as in a
millipede or annelid worm)  but didn't get a good word to make a lujvo
from; pagbu "part" is the best of a bad lot, not implying uniformity of the
segments. pemci (X1 is a poem about x2 author x3 audience x4) conforms
nicely to the pattern for works of art, and I'm tempted to suggest a place
for the stanzas, but it would spoil the pattern.

Returning to the original question, I think kargau (kalri gasnu, open) and
ga'ogau (ganlo gasnu, close) are reasonable for use with files and other
containers.  However, the definition is "x1 (portal/passage/entrance-way)
is open/closed [not] permitting passage/access to x2 by x3", and you have
to metaphorically stretch a bit to relate this definition to "opening" a
book-type object, which everts the entire content, or even "opening" a wine
bottle (which at least has a definite passageway to the interior, though x3
does not traverse it).

James F. Carter          Voice 310 825 2897    FAX 310 206 6673
UCLA-Mathnet;  6115 MSA; 405 Hilgard Ave.; Los Angeles, CA, USA 90095-1555
Email: jimc@math.ucla.edu  http://www.math.ucla.edu/~jimc (q.v. for PGP key)