Baslinux mailing list

[Lee Forrest <lforrestster AT gmail.com>]

On Thu, Jan 11, 2007 at 09:00:44AM -0500, Anthony Albert wrote:

> On 10 Jan 2007 at 15:52, Lee Forrest wrote:
>
> >I've come to the conclusion that the only way to make the tiny
> >linux concept work (10 floppies or less) is to create a new
> >internet. M$ and the merchants rule the one that exists now.
> >
> >We need one that is designed for tiny OSes and slow
> >connections.
>
> gopher was one of the earliest protocols that
> served this purpose. It still functions, actually,
> though many sites have dropped their support.
> http://en.wikipedia.org/wiki/Gopher_%28protocol%29

Gopher is interesting, but too complex. Like BBSs, it tries to
do too much. They are slow and clumsy. I can't see a single
protocol doing the job. I'm talking about a network, not a
server. Decentralized.

I run a mailserver and you run a webserver and someone else
runs a newsserver and someone else runs a directoryserver...

> lynx, links, and their kin are browsers that still support
> text-only browsing, and in many cases, they do a good job...
> if the JavaScript support was just a bit better, they'd be
> perfect, IMO.

I like links a lot. And w3m.

I didn't mean to suggest that there shouldn't be images. Just not
videos or sound, which generally require massive files and very
high bandwidths. And no _gratuitious_ images and no images with
embedded text.

> >It would be easy to write simple servers and clients using
> >bash and netcat if you limited the files they sent/received to
> >plain language text, binary, or basic HTML.
>
> HTML is text, it just gets interpreted by your web browser
> before you see it.

It's more than that, usually. HTML often contains (hyper)links
to image files and other HTML documents, often on other servers.
Hyper Text Transer Protocol. It's not plain language text in any
case.

> >By basic HTML, I mean that what you see on the page is exactly
> >what you get when you print it, although it may take several
> >paper pages to do so.
> >
> >In The Beginning :-), back when scientists and academics ran
> >the internet; before the merchants took over, it provided the
> >virtual equivalents of these basic services, using nothing but
> >basic HTML and plain language text and binary files, on very
> >slow connections:
>
> I believe that the "slow connection" was more of a limiting
> factor than anything "the merchants" have done since.

> By supplying faster and faster connections, the amount of use
> by the same number of people increases to fill the availble
> pipe - I've seen this again and again in my workplace. Order
> more bandwidth, and usage again goes up to about the 80% usage
> level within a few months.

Those last two paragraphs seem to contradict each other. I can't
figure out what you are trying to say.

But "slow" connections are only limitations if you are trying to
turn the network into a virtual shopping mall and a substitute
for the boob tube and sound system. A playground for people
seeking mindless entertainment and a place for merchantile
websites stoked to the gills with eye and ear candy trying
to make themselves appear different and superior to their
competitors, which they are not.

All of the functions I listed can work just fine with "slow"
connections:

Post Office  SMTP/POP
Newstand HTTP/FTP
Magazine Stand HTTP/FTP
Library HTTP/FTP
University NNTP/FTP/HTTP/SMTP/POP/IRC
Public and Private Forums NNTP/IRC
Telephone IRC  (private channel)
Bulletin Boards NNTP/HTTP/FTP
Art Galleries: HTTP/FTP

> [SNIP]
>
> >Throw in a DNS server and an LDAP server (web directory:
> >search engines suck) and you're there. A new-old internet.
> >Make each box a router (simple enough to restrict it's use to
> >members) and the new internet is complete.
> >
> >That LDAP server would also have links to external web sites
> >and ftp sites and mail servers, etc, that were accessible with
> >tiny linux. People would add any such sites they found.
>
> I'm not sure that LDAP can do the sorts of things you want
> here; it tends to be focused on hierarchical structures. I
> think, rather, you're looking for something more along the
> lines of a wiki...

An on-line encyclopedia is a good idea, but that's not what I'm
talking about.

I'm looking for a net directory, something my client can connect
to to find the address of the org or individual I'm looking for.
Not the information, but where to find it. DNS/Seach Engine
combined. A hierarchical structure is perfect.

Something along the lines of the Dewey Decimal System would
work. If I want to find information on pruning roses, then I'd
enter a search string in my client which would would access te
directoryserver and return the addresses and _brief_ descriptions
of the sites:

/science/botany/agriculture//ornamental/flowering/shrubs/roses/cultivation/
/pruning/

None of the search engine nonsense where you get a return on some
stray sentence tucked in a porno page to make it appear in an
unrelated search. Or a page with a poem that includes the phrase
"grows like a rose" shows up in a search for information about
cultivating roses.

A website/page, for example, would register with a description
that would give it a listing in one or more categories. If it
wasn't about pruning domestic roses it would not appear as a
listing in that category.

The 'white pages' would just be listings of the addresses of people
and orgs, by name, probably alphabetically.

Lee

-- 
BasicLinux: Small is Beautiful
http://www.basiclinux.com.ru