test please delete

[WTFH]

Morning all

Heading home today to start my delayed holiday. Speaking of delays, I wonder how late the flight will be.

[original PM]

well it's not dark now

so it's dull n overcast in Manchester.

bit blowy but still not 'cold'

[NickFitz]

Morning denizens

Grey out earlier, but the sun has just broken through

Currently trying to explain the difference between typography with known, named typefaces and the lettering drawn by a traditional signwriter to somebody at Twitter (actually at the company, not a random on Twitter). It's weird that there's a whole generation of people who've never known a world where you can't just pick one of thousands of typefaces for your signage, but have to accept one of the four or five styles your local sign-making person knows how to draw.

This was apropos of her posting a photo of a hand-painted sign on an old American warship that's now a museum and speculating on the font. It would of course have been painted by Hank from the paint shop, who had a pattern book of styles of lettering inherited from his father who worked in the paint shop before him.

At least she's English (she used to be at the Grauniad) so the concept of some craft being a bit older than the Apple LaserWriter isn't as completely alien to her as it would be to an American

[DoctorStrangelove]

This evening's epic is "The Monuments Men", with George Clooney et al.

The Monuments Men (2014) - IMDb

It's sort of interesting in a vague sort of vaguely interesting sort of way.
<spoiler>
Someone just got himself shot in a church which livened things up a bit.</spoiler>

As prompted by someone in the review section, I think I'll watch "The Train (1964)" with Burt Lancaster in glorious B&W this evening.

The Train (1964) - IMDb

Same sort of story, rather better executed in practice.

Oddly, von Ryan's Express springs to mind too, with that Francis Sinatra chap.

Von Ryan's Express (1965) - IMDb

I think I have both of these.

[NickFitz]

Some guy got shot up the road last night, and the local paper is beside itself with glee at finally having some news to report

He was running away from a police raid on his house at the time, and it sounds like his mate who was also running away accidentally shot him in the leg

[NickFitz]

Some guy got shot up the road last night, and the local paper is beside itself with glee at finally having some news to report

He was running away from a police raid on his house at the time, and it sounds like his mate who was also running away accidentally shot him in the leg

Ah, no, looks like that was the paper getting the wrong end of the stick: from the police's official statement on the matter, it appears he was running away from "a business premises" and shot himself in the leg

[NickFitz]

Bacon butty for lunch

Wholemeal bread too, so that's basically health food

[DoctorStrangelove]

Just found a reference to "Harold the bastard" who used to work for Siliconix back in the day.

[covbob]

Time to head back over the Pennines.

[NickFitz]

Dug out some code I haven't touched in getting on for four years, for doing natural language analysis of text to identify possible place names - nothing very clever on my end; it's using a Python library for text analysis. After a few hours in the debugger it's now updated from Python 2.7 to 3.6, and from Django 1.6 to 2.1.

It's not bad, but there's another natural language analysis library for Python which I need to bring in now so I can compare their results with one another

Most of the breakage in my code was in the Epub parser (which, were I to write it now, I would probably do a bit differently). But it was just trivial stuff like updating it to use Python 3's own path traversal library and so on.

But the whole thing eventually managed to download Three Men in a Boat from Project Gutenberg, extracted the text of the book, and identified most or all of the place names therein

I want to be able to run it as a Lambda function eventually - have a queue of books to be analysed, which can be saved to S3 as they're downloaded from PG, which will in turn trigger the Lambda function to do an analysis of the full text and send the resulting possible matches to a web app where I can review them at leisure to weed out the false positives