nemo 🐦 🦣 🐳 🃏
@nemo20000.bsky.social
Senior software engineer – graphics, print and font technologies; PIA/GATF award winner.
RISC OS bore: https://20000.org
RISC OS bore: https://20000.org
Here’s !Teletext doing the correct thing:
[It shouldn’t be remarkable when software does the correct thing but in RISC OS, on this subject at least, it is]
[It shouldn’t be remarkable when software does the correct thing but in RISC OS, on this subject at least, it is]
November 16, 2025 at 5:44 PM
Here’s !Teletext doing the correct thing:
[It shouldn’t be remarkable when software does the correct thing but in RISC OS, on this subject at least, it is]
[It shouldn’t be remarkable when software does the correct thing but in RISC OS, on this subject at least, it is]
Here’s !UniChrs using StyledText to send a Unicode “keypress” to Zap, which ignores it and hence the character is sent as a UTF-8 sequence via ProcessKey which, as you can see, is totally misleading.
![A Cerilica_StyledText message is sent containing an Encoding of “UTF8” [sic] and the UTF-8 sequence of the desired character. The recipient ignores that message, so the sequence is sent instead as individual keypress events, which are indistinguishable from “times sterling” in Latin1.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:iimd7g74rssohysfgvdvgkmn/bafkreiaduokw7gvfl26q5wdjqoslyhlxvhtftjzpybpr2i4tnuhmddng6a@jpeg)
November 16, 2025 at 5:33 PM
Here’s !UniChrs using StyledText to send a Unicode “keypress” to Zap, which ignores it and hence the character is sent as a UTF-8 sequence via ProcessKey which, as you can see, is totally misleading.
The Cerilica_StyledText message can be used to send text with an unambiguous meaning by including an Encoding in the “font name” (which does not have to include a font identifier):
November 16, 2025 at 5:26 PM
The Cerilica_StyledText message can be used to send text with an unambiguous meaning by including an Encoding in the “font name” (which does not have to include a font identifier):
I see the new version of the versatile Amazon Echo is specially optimised for listening to Ozric Tentacles.
That's a Wynne-Win for me.
That's a Wynne-Win for me.
November 13, 2025 at 2:15 PM
I see the new version of the versatile Amazon Echo is specially optimised for listening to Ozric Tentacles.
That's a Wynne-Win for me.
That's a Wynne-Win for me.
I’ve mentioned it before but it continues to intrigue me – compare the mysterious Mexico [sic] alphabet from the M128 International ROM with the CodePage-437 character set.
Both of these are RISC OS Alphabets (but not 7bit numbers) and make a distinction between semantic and appearance Unicodes:
Both of these are RISC OS Alphabets (but not 7bit numbers) and make a distinction between semantic and appearance Unicodes:
November 11, 2025 at 4:24 PM
I’ve mentioned it before but it continues to intrigue me – compare the mysterious Mexico [sic] alphabet from the M128 International ROM with the CodePage-437 character set.
Both of these are RISC OS Alphabets (but not 7bit numbers) and make a distinction between semantic and appearance Unicodes:
Both of these are RISC OS Alphabets (but not 7bit numbers) and make a distinction between semantic and appearance Unicodes:
Scribbly weirdness.
Firefox focus on Android:
Firefox focus on Android:
November 11, 2025 at 9:23 AM
Scribbly weirdness.
Firefox focus on Android:
Firefox focus on Android:
Hang on, the iconbar in Arthur wasn't "fake". It wasn't built into WindowManager as handle -1, but WAS an established convention epitomised by the "Desktop" and replicated by lots of Archimedes programs. For example, here's Highnote, which I used extensively before RISC OS:
November 11, 2025 at 8:40 AM
Hang on, the iconbar in Arthur wasn't "fake". It wasn't built into WindowManager as handle -1, but WAS an established convention epitomised by the "Desktop" and replicated by lots of Archimedes programs. For example, here's Highnote, which I used extensively before RISC OS:
Okudagrams are old-hat, so of course must be supported in RISC OS.
November 10, 2025 at 7:53 PM
Okudagrams are old-hat, so of course must be supported in RISC OS.
Les Dennis is counting the days to the inevitable dramatisation, isn't he.
November 7, 2025 at 3:49 PM
Les Dennis is counting the days to the inevitable dramatisation, isn't he.
Android app menu in UK.
"Don't recommend channel" often needs a full quit/restart or the cached list will still have other content from the same account.
"Don't recommend channel" often needs a full quit/restart or the cached list will still have other content from the same account.
November 7, 2025 at 1:40 PM
Android app menu in UK.
"Don't recommend channel" often needs a full quit/restart or the cached list will still have other content from the same account.
"Don't recommend channel" often needs a full quit/restart or the cached list will still have other content from the same account.
Yes, though I wouldn’t be building any data into nemoBasic – these are separate modules accessed via Service_International.
As you know, the base data is over 2MB, so the trick is to only load the bits you need. As it happens, Canonical Combining Class is titchy:
As you know, the base data is over 2MB, so the trick is to only load the bits you need. As it happens, Canonical Combining Class is titchy:
November 4, 2025 at 2:23 PM
Yes, though I wouldn’t be building any data into nemoBasic – these are separate modules accessed via Service_International.
As you know, the base data is over 2MB, so the trick is to only load the bits you need. As it happens, Canonical Combining Class is titchy:
As you know, the base data is over 2MB, so the trick is to only load the bits you need. As it happens, Canonical Combining Class is titchy:
There’s a lot going on here!
November 4, 2025 at 2:06 PM
There’s a lot going on here!
And I’ve been planning to add magic values for Unicode separator classes such as whitespace, word separators or grapheme clusters (ie “characters”), but that’s heavily data-driven which I’ve avoided in nemoBasic. I think it’s only the EXP function so far that reaches out to a Unicode module:
November 4, 2025 at 1:52 PM
And I’ve been planning to add magic values for Unicode separator classes such as whitespace, word separators or grapheme clusters (ie “characters”), but that’s heavily data-driven which I’ve avoided in nemoBasic. I think it’s only the EXP function so far that reaches out to a Unicode module:
An empty separator splits the string up into “characters”, which in Unicode String Recognition mode means “graphemes” of course:
November 4, 2025 at 1:52 PM
An empty separator splits the string up into “characters”, which in Unicode String Recognition mode means “graphemes” of course:
BBC Basic is not Perl, but nemoBasic’s long strings and DIV$ operator do make it good at consuming and parsing text files.
I’ve just added a long-planned feature: “StrongTabs” seps mode – which isn’t limited to tab characters or one-character separators, but consumes multiple separators as one...
I’ve just added a long-planned feature: “StrongTabs” seps mode – which isn’t limited to tab characters or one-character separators, but consumes multiple separators as one...
November 4, 2025 at 1:52 PM
BBC Basic is not Perl, but nemoBasic’s long strings and DIV$ operator do make it good at consuming and parsing text files.
I’ve just added a long-planned feature: “StrongTabs” seps mode – which isn’t limited to tab characters or one-character separators, but consumes multiple separators as one...
I’ve just added a long-planned feature: “StrongTabs” seps mode – which isn’t limited to tab characters or one-character separators, but consumes multiple separators as one...
The multi-page HSBC advert in today's Times is I think the most complicated bit of web binding/gathering I've ever seen in newsprint. Amazing use of narrowed webs! Very clever.
There was a time this would have been prohibitively expensive, but print advertising is dying.
Omnicom Media is this you?
There was a time this would have been prohibitively expensive, but print advertising is dying.
Omnicom Media is this you?
November 3, 2025 at 12:27 PM
The multi-page HSBC advert in today's Times is I think the most complicated bit of web binding/gathering I've ever seen in newsprint. Amazing use of narrowed webs! Very clever.
There was a time this would have been prohibitively expensive, but print advertising is dying.
Omnicom Media is this you?
There was a time this would have been prohibitively expensive, but print advertising is dying.
Omnicom Media is this you?
I'm not saying it's the flexicable...
But it's the flexicable.
But it's the flexicable.
November 2, 2025 at 7:46 PM
I'm not saying it's the flexicable...
But it's the flexicable.
But it's the flexicable.
The syntax might even allow some emulation – this is just showing off:
October 29, 2025 at 3:21 PM
The syntax might even allow some emulation – this is just showing off:
BBC Basic is available for multiple platforms and processors, but there’s one keyword (well, two) that’s definitely not portable – CALL.
I decided to address that (ha!) in nemoBasic:
I decided to address that (ha!) in nemoBasic:
October 29, 2025 at 3:21 PM
BBC Basic is available for multiple platforms and processors, but there’s one keyword (well, two) that’s definitely not portable – CALL.
I decided to address that (ha!) in nemoBasic:
I decided to address that (ha!) in nemoBasic:
Great minds and fools...
October 29, 2025 at 12:18 PM
Great minds and fools...
Mr Kipling, this is pathetic. Dreadful. Abject.
I almost began to consider not eating it.
It actually looks better in the photo. Damn you Google camera wizards.
I almost began to consider not eating it.
It actually looks better in the photo. Damn you Google camera wizards.
October 28, 2025 at 2:26 PM
Mr Kipling, this is pathetic. Dreadful. Abject.
I almost began to consider not eating it.
It actually looks better in the photo. Damn you Google camera wizards.
I almost began to consider not eating it.
It actually looks better in the photo. Damn you Google camera wizards.