Re: Unicode is universal, so how come that universality doesn’t apply to digits?

[Zach Lym]

On Mon, Dec 21, 2020 at 6:08 PM Ken Whistler <kenwhistler at sonic.net> wrote:
> On 12/21/2020 5:00 PM, Zach Lym via Unicode wrote:
> PDFs do not rank well on Google, you can't deeply link to specific
> sections,
>
> Actually, you can, if you set them up correctly:
>
> https://www.unicode.org/versions/Unicode13.0.0/ch22.pdf#G12146
>
> That links right to Table 22-3, Script-Specific Decimal Digits on p. 829, in Section 22.3 of the latest version of the core specification.

I don't want to be rude ... but protips just enable user abuse 😼.
Being an expert in something insulates you from the harsh realities
faced by your users 😔.

During my review of filename normalization decisions ... experts were
confused and made poor choices at virtually every step.  Usability
engineering views confused end-users as a signal that important
information isn't being surfaced in an appropriate manner.  When you
are failing to meet your target demographic of
smart-people-in-a-hurry, what hope is there for us idiots?

Not much, **because no one reads manuals.**  That is one of the most
reliable findings of technical documentation research stretching back
to 1987 [1].

>  and they are generally a PITA to work with.
>
> Well, your mileage may vary. HTML has its own PITA aspects.

Everything involves trade-offs, but Unicode's PDFs are even worse than
the IETF's typewriter emulator 😅.

>   The Unicode
> standard publishes PDFs *not* because it is a good idea, but because
> it's inconvenient to change a 30-year-old publishing workflow.
>
> 20, not 30, actually. Prior to Unicode 3.0, the Unicode Standard was done with a different family of editorial tooling.

So what are you using, DocBook?

> But yeah, it is inconvenient to change, especially since the document is riddled with hand-tweaked figures and hacked up fonts. And it's a thousand pages long, and it has internal indexing and the sections, figures, and tables are all cross-referenced in the document. And oh, did I mention? It's a thousand pages long.

Oh, no.  That sounds terrible ... for someone who isn't a print 🥷!

My father taught graphic design, so I grew up messing around with
PageMaker and doing table based HTML layouts.  My post high-school job
involved Vietmanese and Amharic print work.  The Ethiopian history
books weren't a thousand pages long, but they had reference indexes,
figures, and codepage mixing ... the whole nine 😸.

> Various folks have wanted to reformat it to something more web-friendly and searchable over the years,

I only suggested additional cross-references within the standard and
possibly a new technical report, which would be more *user* friendly.

The UX rabbit hole started based on my assertion that the
disproportionate amount of effort put into the identifier
documentation and widespread support for i18n variable names is more
than *just* a correlation.

> but they have tended to discover other things that they needed to do when faced with the actual amount of work involved. ;-)

That is not something an outsider could do, as the primary audience
for any product are the people who make it.  And if the "insiders"
don't see a problem....

Hence my invocation of Conway's law: the standard reflects the
particular bureaucratic mould in which it is formed.

ツ
- Zach Lym
[1]: https://scholar.google.com/scholar?hl=en&as_sdt=0%2C38&q=minimal+manual&btnG=