Say you work as a professional, and your professional body sends around a form for you to fill out. They tell you it will be published on the web, say to help people find their nearest qualified aerospace engineer, of which you are one of course. Now, if you fill in an email address on the form and it gets published you will get emails. It is surprising how many people think this will not happen.

Which brings me to what my real point is: a whole load of data that gets thrown out there on the web is published without considering either its intention or its purpose.

If you make your phone number public, is it decipherable? The bottom line is that I don’t want phone numbers like (0)-555/767.768/9 appearing anywhere. [OK, below is a long post-amble about what I think a well formed number might be. Read it to see how finikity (which may or may not be a word, making it all the more embiggening) I am.]

If you write your number like this people will probably still be able to reach you, but you’re making it hard for them, and if you’re inputting the number on the web somewhere, you’re probably caused a few people to scratch their heads and wonder if they should bother coding solutions for numbers of such opacity. (They should not.)

In case you’re getting angry at coders at this stage, the flip-side is the dumb entry field that insists you enter your number in some bizarre format chosen by a coder through ignorance/laziness. (Yes Bord Gáis a + is valid in a telephone number, in fact it is the single best starting character for a number to have, but enough about that… )

Where was all this going?

Oh yes, why is all this information going on the web in the first place?

A lot of official data now seems to find a place on the web in PDF format. The motivations behind this may be laudable; the original may be a Microsoft Word document and PDF is a more open format. However, PDFs tend to be designed for human rather that machine legibility and lack the semantic structures that are increasingly available (if underused) in html/xml formats.

If the official Canadian list of area dialling codes is in a PDF that lacks a readable table structure then someone (or worse still, more than one person) will create unofficial lists using wikis or web pages. These will only maintain some level of compatibility with each other and the original source, so any further sources of information that verify themselves against these pages become less and less reliable. This is all a long way from Tim Berners-Lee’s vision for a semantic web.

If governments are responsible for making data they hold accessible, and this is good in any number of ways, then they also have a responsibility for the level of accessibility. A locked filing cabinet in a basement that is accessible by asking the staff member responsible, available the first Monday of the month, between 9:00 and 9:30am, is a level of accessibility…

This is not a purely esoteric complaint. Somebody pays for that list of dialling codes, and those emergency service numbers and a whole host of data that has been decreed web worthy, but very little thought seems to go into how much of that information appears to the web or how its value could be maximized.

ISO, the International Organisation for Standards publish a lot of information on the web. Take their currency pages for example, which have nice tables of data. Well, their HTML is invalid; the doctype does not match the data contained. Their entity names are arbitrarily chosen forms of country names, even though they themselves define unique identifiers for countries, so the table must be interpreted for it to have value. I guess if ISO can’t get this right there is little hope for other folks.

Is there a point to all this? Well the ISO pages are better than a PDF, and PDFs are better than MS Word documents, and all are better than no information at all. There is at least a chance that something like linked data will be widely used and provide a more useful web, but, in the meantime, just a little thought about the quality of information and reasons for it to be placed on the web will make me a lot happier…

*Post-amble

What does a dash (-) in a phone number mean, why was it put there? Probably it was just put in to separate the number into human digestible chunks 555-767-768. Whitespace would have been as good but no problems so far.

What does a slash (/, virgule) mean in a phone number? In a sentence it might have a dash type meaning but I would usually use it to designate a substitution or a logical OR type statement. I would avoid using a – for OR and also avoid using a / for a join.

The number 555-767-768/9 then means that either 555-767-768 or 555-767-769 will get you through.

The number 555/767/768-9 would be ambiguous to me: it could mean the same as above but I might try 555767768 wait then press 9, or some such.

The use of a full-stop as a visual separator seems dubious; it has other meanings both in number sequences and in textual contexts.