Re: [ossig] Nothing exciting in Comp.Sci. anymore?

[Mukhsein Johari <pythonium@xxxxxxxxx>]

--- Imran William Smith <iwsmith@mimos.my> wrote:
> Dinesh Nair wrote:
> > On Tue, 17 Feb 2004, Mukhsein Johari wrote:
> > 
> > 
> >>So, now, when you learn about computer science,
> you learn about the file
> >>as if it were an element of nature, like a photon.
> That's a dangerous
> >>mentality. 
> > 
> > rubbish !
> > 
> > files are an incarnation of a data structure. if
> he dislikes them so much,
> > then perhaps he should have presented another
> concept or methodology.
> > 

Hmm..I'm not willing to rubbish it out of hand,
though. Most people don't think about files - it's
just there. Just as most people don't think about
phone numbers. It's a product of human progress.

It's not the only way to do things, though. 

A file is simply an EOF terminated sequence of bytes.
A 1-dimensional structure - you only have base and
offset (how far from the base) to read and write data.

Is 1-dimensional storage the only way? No, you could
have 2 dimensional storage for example, using
O(origin), x, y
Which would make reading, writing and processing
certain kinds of data more efficient (eg. tables,
pictures)

You could have tree-style storage (xml is an example
of a serialized tree - note the need to parse xml. ie.
transform a 1 long line structure into a tree
structure in memory) or even network-style storage.
Each node pointing to one or more other nodes - which
could be more fault tolerant for example, corruption
in one node does not make the whole structure
unusable. (think internet).

Using conventional files to store data leads to the
problem of file formats - the way in which data is
understood by applications is dependent on how the
data is parsed. (the nature of files makes parsing
implicitly required)

Eg. open an html file with a normal text editor will
yield different results to opening it in a web
browser.

It also does not have any concept of metadata - this
has to be added on by the FS (if at all - like file
owner etc.). Various methods like extended attributes
and so on have been suggested and used to address
this.

What I understand him to say is that we should not let
our 'knowledge' of what _is_ limit our thinking in
creating what _could_ be. Lateral thinking and
thinking out of the box will likely result in vastly
different solutions that may be far superior to
current convention. 

eg. Instead of upgrading a server with more ram, more
disk, more bandwidth to support more users, you could
let the users themselves do the 'upgrading' -
peer-to-peer file-sharing was born!



__________________________________
Do you Yahoo!?
Yahoo! Mail SpamGuard - Read only the mail you want.
http://antispam.yahoo.com/tools


---------------------------------------------------------
To unsubscribe: send mail to ossig-request@mncc.com.my
with "unsubscribe ossig" in the body of the message