酷URI永不变：Web架构师Tim Berners-Lee谈URI设计原则

Cool URIs don't change

What makes a cool URI?

A cool URI is one which does not change.

What sorts of URI change?

URIs don't change: people change them.

There are no reasons at all in theory for people to change URIs (or stop maintaining documents), but millions of reasons in practice.

In theory, the domain name space owner owns the domain name space and therefore all URIs in it. Except insolvency, nothing prevents the domain name owner from keeping the name. And in theory the URI space under your domain name is totally under your control, so you can make it as stable as you like. Pretty much the only good reason for a document to disappear from the Web is that the company which owned the domain name went out of business or can no longer afford to keep the server running. Then why are there so many dangling links in the world? Part of it is just lack of forethought. Here are some reasons you hear out there:

We just reorganized our website to make it better.

Do you really feel that the old URIs cannot be kept running? If so, you chose them very badly. Think of your new ones so that you will be able to keep then running after the next redesign.

We have so much material that we can't keep track of what is out of date and what is confidential and what is valid and so we thought we'd better just turn the whole lot off.

That I can sympathize with - the W3C went through a period like that, when we had to carefully sift archival material for confidentiality before making the archives public. The solution is forethought - make sure you capture with every document its acceptable distribution, its creation date and ideally its expiry date. Keep this metadata.

Well, we found we had to move the files...

This is one of the lamest excuses. A lot of people don't know that servers such as Apache give you a lot of control over a flexible relationship between the URI of an object and where a file which represents it actually is in a file system. Think of the URI space as an abstract space, perfectly organized. Then, make a mapping onto whatever reality you actually use to implement it. Then, tell your server. You can even write bits of your server to make it just right.

John doesn't maintain that file any more, Jane does.

Whatever was that URI doing with John's name in it? It was in his directory? I see.

We used to use a cgi script for this and now we use a binary program.

There is a crazy notion that pages produced by scripts have to be located in a "cgibin" or "cgi" area. This is exposing the mechanism of how you run your server. You change the mechanism (even keeping the content the same ) and whoops - all your URIs change.

For example, take the National Science Foundation:

NSF Online Documents

http://www.nsf.gov/cgi-bin/pubsys/browser/odbrowse.pl

the main page for starting to look for documents, is clearly not going to be something to trust to being there in a few years. "cgi-bin" and "oldbrowse" and ".pl" all point to bits of how-we-do-it-now. By contrast, if you use the page to find a document, you get first an equally bad

Report of Working Group on Cryptology and Coding Theory

http://www.nsf.gov/cgi-bin/getpub?nsf9814

for the document's index page, but the html document itself by contrast is very much better:

http://www.nsf.gov/pubs/1998/nsf9814/nsf9814.htm

Looking at this one, the "pubs/1998" header is going to give any future archive service a good clue that the old 1998 document classification scheme is in progress. Though in 2098 the document numbers might look different, I can imagine this URI still being valid, and the NSF or whatever carries on the archive not being at all embarrassed about it.

I didn't think URLs have to be persistent - that was URNs.

This is the probably one of the worst side-effects of the URN discussions. Some seem to think that because there is research about namespaces which will be more persistent, that they can be as lax about dangling links as they like as "URNs will fix all that". If you are one of these folks, then allow me to disillusion you.

Most URN schemes I have seen look something like an authority ID followed by either a date and a string you choose, or just a string you choose. This looks very like an HTTP URI. In other words, if you think your organization will be capable of creating URNs which will last, then prove it by doing it now and using them for your HTTP URIs. There is nothing about HTTP which makes your URIs unstable. It is your organization. Make a database which maps document URN to current filename, and let the web server use that to actually retrieve files.

If you have gotten to this point, then unless you have the time and money and contacts to get some software design done, then you might claim the next excuse:

We would like to, but we just don't have the right tools.

Now here is one I can sympathize with. I agree entirely. What you need to do is to have the web server look up a persistent URI in an instant and return the file, wherever your current crazy file system has it stored away at the moment. You would like to be able to store the URI in the file as a check, and constantly keep the database in tune with actuality. You'd like to store the relationships between different versions and translations of the same document, and you'd like to keep an independent record of the checksum to provide a guard against file corruption by accidental error. And web servers just don't come out of the box with these features. When you want to create a new document, your editor asks you for a URI instead of telling you.

You need to be able to change things like ownership, access, archive level security level, and so on, of a document in the URI space without changing the URI.

Too bad. But we'll get there. At W3C we use Jigedit functionality (Jigsaw server used for editing) which does track versions, and we are experimenting with document creation scripts. If you make tools, servers and clients, take note!

This is an outstanding reason, which applies for example to many W3C pages including this one: so do what I say, not what I do.

Why should I care?

When you change a URI on your server, you can never completely tell who will have links to the old URI. They might have made links from regular web pages. They might have bookmarked your page. They might have scrawled the URI in the margin of a letter to a friend.

When someone follows a link and it breaks, they generally lose confidence in the owner of the server. They also are frustrated - emotionally and practically from accomplishing their goal.

Enough people complain all the time about dangling links that I hope the damage is obvious. I hope it also obvious that the reputation damage is to the maintainer of the server whose document vanished.

So what should I do? Designing URIs

It is the the duty of a Webmaster to allocate URIs which you will be able to stand by in 2 years, in 20 years, in 200 years. This needs thought, and organization, and commitment.

URIs change when there is some information in them which changes. It is critical how you design them. (What, design a URI? I have to design URIs? Yes, you have to think about it.). Designing mostly means leaving information out.

The creation date of the document - the date the URI is issued - is one thing which will not change. It is very useful for separating requests which use a new system from those which use an old system. That is one thing with which it is good to start a URI. If a document is in any way dated, even though it will be of interest for generations, then the date is a good starter.

The only exception is a page which is deliberately a "latest" page for, for example, the whole organization or a large part of it.

http://www.pathfinder.com/money/moneydaily/latest/