[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Why HTTP URL's?
- To: srfi-84@xxxxxxxxxxxxxxxxx
- Subject: Why HTTP URL's?
- From: Andrew Wilcox <awilcox@xxxxxxxxxxxxxxxxx>
- Date: Thu, 26 Jan 2006 17:08:26 -0500
- Delivered-to: srfi-84@xxxxxxxxxxxxxxxxx
- Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:sender:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=hsIyNHppj2DEagxia8Lg8/e5j/oKsLAwdBmT0Z2C/2P8k6JCow8YcV6tAuaAfLBC8M5nz7Um0m2Yz4xwkGwk7fFjLUMGNTki024VYheZIeGUuFybBIOQxbkTeEGDO0TUqbl3BKBCMckWgPbqtkQDkwP56ULgPQBTAGxfVbMjFO4=
- Sender: andrew.wilcox@xxxxxxxxx
Neil Van Dyke writes:
> If "http"-scheme URLs were used for identifying libraries, I'd think
> that each URL would actually be a reference to a Web object that
> could by retrieved via HTTP to define the referenced library. If we
> only wanted a name for a library, I'd suggest not using URLs. Is
> that what you had in mind?
Ray Blaak writes:
> If URLs are allowed, why restrict to only http:. Why shouldn't ftp:
> be allowed, or jar: to extract contents from an archive. Why not
> file: for local libs, or for during development?
My goal is to allow a programmer to easily and quickly choose a
readable name for their library that is unlikely to conflict with
other library names.
For programmers who are not putting their library under schemers.org
or who have their own domain name, I was looking for something that
many people would have, or could get easily, that would be readable
and would be largely unique to them.
Someone's email address would, I think, be a good example. They are
more or less readable, and most people have one or can easily get one.
While an email address can change hands, it is usually fairly unique
to the individual. However I didn't think that people would want to
use their email address because of spam considerations.
Most people have their own web page, or can easily get one, and a web
page address is easily recognizable.
A clear disadvantage, I think, of using a web page address as an
identifier is that people will then expect that they should be able to
go to that address and get something. The library itself perhaps. Or
at least meta data about it. And if you can do that, you should be
able to do it for open-source- library based names. Or for
domain-name based names. Or allow the meta-data to be pulled from ftp
or jars or files.
Using web page addresses as a name is not my first choice, but I do
think it is important to have it as an option.
Maybe leaving off the "http://" would make it look less like a URL and
more like a name, and perhaps help avoid the confusing expectation
that you should be able to go to that web page and get something.