Search This Blog

Sunday 29 November 2009

Word Puzzles

While writing my last post I noticed that the spellchecker didn't recognise the word 'aerogramme', which is fair enough, as the whole point of the post was that aerogrammes are no longer really needed. But just now I was writing a letter in Google Docs, and it put the red, wiggly lines under 'blogging' and 'Google', both words you'd think it might have heard of.

I don't know where Google Docs keeps its dictionary. Maybe it piggybacks on one that's already on my PC,  though Blogger, in which I'm writing this and which is also owned by Google, is quite happy with both words. (But not with 'Blogger'! The plot thickens.)

Perhaps this reflects some streak of self-effacement among Google developers. I don't know. I was also intrigued a few years back when I spotted that Microsoft Word didn't recognise the word 'Dilbert', but was quite happy with 'Dogbert'. Presumably someone on the Word team was a member of Dogbert's New Ruling Class, and snuck that one in. Considering that the Excel team once hid a flight simulator in their program, one extra line in a dictionary file wouldn't have been too hard to manage.

And now it's time to check again whether Haxby has had its water supply reconnected, following the 'incident' (as Yorkshire Water usefully describe it on their web site) earlier today. I do miss the 21st century.

Another Superseded Technology

Rummaging through our stationery drawer this morning, looking for a suitably sized envelope, it occurred to me that most of the difference between a tidy drawer and what we actually have was the pile of old aerogrammes at the back. Actually two piles, I suppose: the one that I'd bought to our marriage, and the one that Julie had. We've been living together for over 15 years now, but I doubt if we've used a single aerogramme in all that time.

For those of you too young to remember aerogrammes (which seems to include the Blogger spellchecker), they are (were?) very thin pieces of blue letter paper, carefully shaped with gummed edges that allowed the aerogramme to be folded in three to form its own envelope, which would then be posted to exotic destinations around the world. The light weight of the paper, and the prohibition on any enclosures, made them ideal for air mail.

I do have friends in foreign parts, but, apart from the annual Xmas card, anything I send to them goes over the internet nowadays. I don't even know how well an aerogramme would go through a laser printer (okay, I do now: not too badly, except you need a larger top margin).

Most of our aerogrammes are postage prepaid, which will make parting with them a little painful, but the hard truth is that we'll get more utility out of them by freeing up some space in a drawer than by leaving them there untouched for another fifteen years.

Putting them into the recycling bin (where else?), I notice that on one of the packs there's a 50th Anniversary message (1943 - 1993): "50 years ago, when Churchill was Prime Minister, Aerogrammes took off as a vital means of communication." It finishes off with, "Aerogrammes - still the easy way to keep in touch."

Sunday 22 November 2009

How difficult is is to sync a phone and a PC?

As I've mentioned before, my mobile is a Nokia N85 smartphone. This is a good device, though I wish I'd known how limiting texting was before I bought something that didn't have either a keyboard or a stylus. Where it falls down is in its communication with my PC. The Nokia software that comes with the phone allows you to synchronise contacts and calendar data with Outlook, Outlook Express, and Lotus Notes. I use Mozilla Thunderbird.

Not to worry. There's a program you can buy called Mobile Master, which lets you synchronise various mobile phones with sundry mail clients. Problem solved.

Well, not quite. Right from the start I've been suffering all sorts of glitches with the data transfer. Some events that recur every fortnight in Thunderbird show up every week on the phone, others shift by a day sometimes, and the 'Other data' field in the Contacts gets strangely mangled when sent back from phone to PC. All that was insignificant compared to the way that the entire calendar got downloaded to the phone twice (most times; tantalisingly, it would work occasionally).

There are fairly frequent updates to Mobile Master, which I would install hopefully, and a few months after I'd bought the phone, suddenly the transfer started to work every time. Events would get sent down just the once, and the other misbehaviour I fixed by deleting the affected items in Thunderbird and recreating them. (No idea why this was necessary, as they were fine when I used to synchronise with an iPaq.)

I stopped updating Mobile Master then. The newer versions were mostly concerned with supporting the latest models of mobile phone. But then, a couple of weeks ago, I took leave of my senses and upgraded to 7.5.5.

It was mostly okay, if you weren't bothered about not having any calendar items on your phone any more.

I raised a bug report, which had as much effect as the one I'd raised about getting stuff sent down twice. As before, I got back an automated reply with the ambiguous message:

We have received the following message and will answer as soon as possible.

Sorry, but we can no longer answer questions whether this or that phone is supported.

Had they written off my query as a question about support for the n85, or was that just a standard line added to every reply? As I've never received a further reply, the question will remain open.

Luckily I still had an earlier version of their installer on my hard drive, so I was able to go back to that. A bit too early, unfortunately, as I'm back to having two copies of the calendar again.

Some time soon we are promised Thunderbird 3.0 will be released. Unlike Thunderbird 2, this will have a calendar system in-built, instead of as one of two possible add-ins. Perhaps it will build up enough market share then to make Nokia think it's worth supporting. (A quick web search suggests that Thunderbird 2 only has 1.12% market share at the moment, so I might be whistling in the dark on this. On the other hand, Lotus Notes only has 1.72%.)

In the meantime, I wait for another update to Mobile Master, and remind myself more often not to bother updating something that already works perfectly well.

Sunday 8 November 2009

Thirty Years of Experience - A Perspective, part 3

I think it's time to finish off this set of posts before the title has to change to 'Thirty-One Years of Experience'.

I started off trying to work out how much use thirty years of programming experience was really worth, given that so much of it was in technologies that are no longer extant. I've come, sadly, to the conclusion that much of the first fifteen years or so can be safely skimmed through by anyone reading my CV. In the field of Software, specific knowledge of languages, libraries and operating systems dates all too quickly. If you're not constantly learning new technologies, you can end up fit only for maintaining legacy code.

And yet I still hold that I've gained something from having programmed since 1980 that I wouldn't have now if I'd only started in 2000. Thirty years of debugging have given me an intuition for tracking down bugs, to the extent that sometimes I can't even explain to myself afterwards how I got to the solution. I believe this is because debugging is quite often a logic problem, and practice makes you better at it irrespective of what language or platform you're using.

Over thirty years I've seen trends come and go, supposedly 'killer' languages have their day and then fall away (Ada, you promised so much!). It would be nice to say that my experience lets me spot which upcoming technologies are the ones to follow, and which are going to fail, but that would be untrue. I write as someone who once learned Forth.

In the end, it's not so important how much my experience is worth. The bottom line is that, after nearly three decades in the field, I still feel as I did when I received my first pay packet back in 1980: that it's a nice life when you get paid to do something you'd willingly do for a hobby.

Saturday 7 November 2009

More lessons in XHTML

So, following on from my last blog, I've now done the research I mentioned, and found out that the way to deal with ampersands in URLs is to write them as XML entities, i.e. '&', which is slightly embarrassing, as that should have occurred to me given that I've been using XML for nearly a decade.

And the XHTML-approved way of making a hyperlink open a new window is to replace the 'target' attribute with an 'onclick' attribute which calls the relevant JavaScript function to do it.

Something else I've just learned, having just previewed this blog, is that typing '&' out in full is more than a bit dumb, as the browser, of course, renders it as '&'; time for a sneaky trick to fool the browser. One last question: how can a modern browser possibly not know that 'blog' is a proper word?

The Quest for XHTML Perfection

This afternoon I remembered the existence of the W3C Markup Validation Service, http://validator.w3.org/, and decided to submit the web sites I maintain. As usual I was appalled by the number of omissions, typing errors, and plain mistakes that your average browser will quite happily work round without telling you. Unterminated paragraph blocks, lists inside blocks, attribute values with a quote at one end but not the other, to name just three blunders which now no longer afflict my HTML.

The ideal is to press the Validate button and get back the message, "This document was successfully checked as XHTML 1.0 Strict!" Once you've got that, you're allowed to put a W3C badge on the page to tell the world how compliant you are, or at least the 0.1% or less of the world who've ever even heard of the W3C, or XHTML. Rather than confuse the other 99.9% of the world, I have decided not to use this badge. Also, I would then live in fear of inadvertently invalidating the page through some stupid edit, and incurring the wrath of the W3C.

Another reason for not boasting of my pages' compliance is that not all of them are, and there doesn't seem to be much I can do about it. At one of my sites, gratuitous link to encourage Google to notice it, I include a link to a specific page at the UK Charity Commission's web site. Even though I'm quoting their URL exactly, the W3C is flagging up errors in the URL and blaming me. This seems a mite pedantic of them. Am I supposed to tidy up someone else's web site?

The link in question is http://www.charity-commission.gov.uk/ShowCharity/RegisterOfCharities/CharityWithoutPartB.aspx?RegisteredCharityNumber=1119272&SubsidiaryNumber=0, and the presence of the ampersand near the end causes W3C to kick out four errors and three warnings. And now, by putting that link into this blog, I've managed to make this page non-compliant too!

Footnote: thinking about my question above, I wondered if I could tidy up the URL. Turns out I can leave the part from the ampersand onwards! I got lucky there, but I think my point still stands.

Now my only non-compliance is the use of the 'target' attribute in hrefs. Need to do a bit of research about why that's not valid XHTML.