💾 Archived View for gemi.dev › gemini-mailing-list › 000391.gmi captured on 2024-05-26 at 15:51:26. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2023-12-28)
-=-=-=-=-=-=-
I'm searching for the like-minded people that agree that URL syntax for hyperlinks means desperately clinging to the mistakes of the past. It seems that in a world of rapidly expanding browser scope some fundamentals are set in stone even for known violators. And the current opposition seems to be all ...return to the Tim Berners-Lee era of the WWW. As if few to none want to go further than attempting a reaction to certain annoyances and privacy encroachments that the Web has, without considering the features the Web never realized. Of course it's somewhat breaking away from a paradigm, but I strongly feel that consistent support for URNs in preference to URLs in electronic documents will further the reach of hypertext beyond the always online crowd, by introducing a layer of indirection that greatly simplifies eliminating single points of failure the URL-cross-referenced documents are infamous for, contributing to the perception that electronic resources must be transient. And I'm not saying it's good to base on a legacy URN specification that obeys RFC3986 either, as evidenced by the mess that magnet: URIs present because of ugly percent encoding.
On Sun, Sep 27, 2020 at 9:06 AM <persist at localhost> (really??) wrote:
Of course it's somewhat breaking away from a paradigm, but I strongly feel
> that consistent support for URNs in preference to URLs in electronic
> documents will further the reach of hypertext beyond the always online
> crowd, by introducing a layer of indirection that greatly simplifies
> eliminating single points of failure the URL-cross-referenced documents are
> infamous for, contributing to the perception that electronic resources must
> be transient.
>
In practice, given the existence of doi.org and similar systems, as well as
URL shorteners (I often use them to generate de facto URNs), and the fact
that HTTP and Gemini both allow protocol-level redirection, I think what we
have is as good as it's going to get. DNS is the most successful and
widespread federated database system in the world, and attempting to
replace it is likely to be a pipe dream. Link rot is a thing, but so is
database rot and protocol rot. As the French proverb has it, "tout passe,
tout lasse, tout casse et tout se remplace" (Everything passes, everything
grows weary, everything breaks, and everything gets replaced).
And I say this as a person who has dealt with SGML/XML public ids and URNs,
and even wrote RFC 3305 to define a mapping from the former to the latter.
It's a dead end. Of course, I'm certainly open to the idea that my
conclusions were premature, but the burden of persuasion is on you. I know
the merits of stable names, you needn't convince me of that. You need at
least a sketch of a system that is clearly superior to URLs as we know them
today.
John Cowan http://vrici.lojban.org/~cowan cowan at ccil.org
Your worships will perhaps be thinking that it is an easy thing
to blow up a dog? [Or] to write a book?
--Don Quixote, Introduction
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.orbitalfox.eu/archives/gemini/attachments/20200927/533f
40d4/attachment.htm>
Agree with John. All that is solid melts into air? John Cowan <cowan at ccil.org> writes: > On Sun, Sep 27, 2020 at 9:06 AM <persist at localhost> (really??) wrote: > > Of course it's somewhat breaking away from a paradigm, but I strongly feel >> that consistent support for URNs in preference to URLs in electronic >> documents will further the reach of hypertext beyond the always online >> crowd, by introducing a layer of indirection that greatly simplifies >> eliminating single points of failure the URL-cross-referenced documents are >> infamous for, contributing to the perception that electronic resources must >> be transient. >> > > In practice, given the existence of doi.org and similar systems, as well as > URL shorteners (I often use them to generate de facto URNs), and the fact > that HTTP and Gemini both allow protocol-level redirection, I think what we > have is as good as it's going to get. DNS is the most successful and > widespread federated database system in the world, and attempting to > replace it is likely to be a pipe dream. Link rot is a thing, but so is > database rot and protocol rot. As the French proverb has it, "tout passe, > tout lasse, tout casse et tout se remplace" (Everything passes, everything > grows weary, everything breaks, and everything gets replaced). > > And I say this as a person who has dealt with SGML/XML public ids and URNs, > and even wrote RFC 3305 to define a mapping from the former to the latter. > It's a dead end. Of course, I'm certainly open to the idea that my > conclusions were premature, but the burden of persuasion is on you. I know > the merits of stable names, you needn't convince me of that. You need at > least a sketch of a system that is clearly superior to URLs as we know them > today. > > > > John Cowan http://vrici.lojban.org/~cowan cowan at ccil.org > Your worships will perhaps be thinking that it is an easy thing > to blow up a dog? [Or] to write a book? > --Don Quixote, Introduction
---
Previous Thread: [ANN] GemGoHaNe -- Gemini Gopher HackerNews Mirror