Statutes and URLs, Part 1

I had occasion recently to work with some URLs for legislation and was struck yet again by the peculiarity of fetching and “citing” statutes in this way when on or linking to the internet. I have some narrow, specific concerns, that I’ll talk about in this post; and then in Part 2 I’ll wander a bit in the land of speculation, to see how else it might be done.

Here and now I want to complain about long, illegible URLs, the kind that represent the raw output of queries to a database. This is not the first time by any means that URLs, long or short, have been talked about on Slaw — which isn’t too surprising, as it’s part of typical lawyerly concern for detail in this time of transition from paper to pixel.

My entry point to the matter was Ontario’s Employment Standards Act, 2000 the online and authoritative version of which is to be found on the Ontario E-Laws site, itself the possessor of a clear, even memorable URL: ; and better still, it works whether you use the “e-laws” portion as a subdomain ( or a subdirectory ( Some good thinking here.

Why, then, does that good thinking not reach all the way down?

The E-laws site offers two ways of getting a statute: searching and browsing. Because the search box comes first on the page, I used it to search for “employment standards act” and came up with 213 hits in 22 documents, finding the link to the Act itself as the number 3 result. (There should be a better way to present search results when an act title is searched for — but that’s another matter.) A click on that link got me to the Act — with the following as the URL: isysquery/d6e80baf-5247-4d5c-83dd -61ca8454fc9b/3/frame/?search=browseStatutes&context=

Now, I’ve seen worse, when it comes to length. But no one could claim that this is anything but the unlovely, unreconstructed product of a database search. It is incomprehensible to the human eye, and the dangling “=” is a cause for anxiety. There is no way to know whether you’ve somehow accidentally curtailed it when copying, because it has no logical structure that is apparent. All in all, this is not something you’d like to work with.

Now, the real URL for the Employment Standards Act, 2000 is obtained by browsing via E-Laws’ alpha lists, which produces this: /english/elaws_statutes_00e41_e.htm

which is, in fact, human-readable, more or less.

What is needed is a way of getting the better URL if you come in via a search. That could be by making the better link appear at the top, the method CanLII uses to provide a “friendly” URL when you land on a judgment as the result of a search. Or, it could be by simply causing the long search link to resolve automagically as the good one. At present, you have to back out and go back in through a “browse.”

In British Columbia, to take another example, there simply is no friendly URL. The URL for the B.C. Employment Standards Act, whether found by searching or browsing, is the following:–%20e%20–/employment%20 standards%20act%20%20rsbc%201996%20%20c.%20 113/00_96113_01.xml#FOUND-NOTHING

I’m particularly fond of the “FOUND-NOTHING” closer, which beats Ontario’s “=”. But the proliferation of %20 takes the cake. %20 is the hex symbol for a space, and there should be no spaces in URLs. Here again, we’re given the the undigested output of a database search and yet another chance to thank our lucky stars for copy and paste.

Federal laws, to take just one more example, do an almost adequate job of serving up decent URLs. Finding the Employment Equity Act by browsing alphabetically renders this URL:

Except for the rejection of dessert at the end, this is almost comprehensible. And it can be (and therefore should have been) reduced to:

But, as with Ontario, the search function is presented front and centre on the federal laws page, so most users will find their statutes using search. If you should get to the Employment Equity Act by this route, you’ll find this URL: command=home&caller=SI&search_type=all&shorttitle=employment%20 equity%20act&day=22&month=8&year=2009&search_domain =cs&showall=L&statuteyear=all&lengthannual=50&length=50&noCookie

All that work and still no cookie!

There is a link on the page to the full document in html, but it gives this nearly identical URL: command=home&caller=SI&search_type=all&shorttitle=employment%20 equity%20act&day=22&month=8&year=2009 &search_domain=cs&showall=L&statuteyear=all&lengthannual =50&length=50&noCookie

One answer to the mess is to turn to CanLII for statute URLs, something I’m normally happy to do, because they’re much friendlier and consistent across jurisdictions. But CanLII isn’t authoritative. And it shouldn’t be the job of an underfunded, small organization to, in effect, clean up after governments.

Why is all of this a problem, anyway? Well, one obvious reason has to do with precision in linking: increasingly legal documents are digital, so references to laws are hyperlinked; one letter wrong, and the URL will break, or take you to a place you don’t want to go; foolishly complex URLs make it more likely that mistakes will occur in linking, then. There is also the reason that any such long URL in print is simply useless as a functioning reference: no one can or would copy out most of the URLs I’ve given here. Then, too, there’s something alienating in having this gobbledygook represent a statute, particularly when there are other, more elegant ways to do it — which I’ll think about in Part 2.


  1. Simon

    In the UK, OPSI have given this plenty of thought and have now developed a system of permanent URIs for legislation based on linked data standards which will be fully implemented on a new Single legislation Service soon. It would be a huge boon if other publishers followed their lead.


  2. Aloha Simon. I’m looking forward to part 2 (and wrote about our thoughts in the area for LexPublica on our blog at )