(2024-05-20) A look from another side of the spectrum... again
--------------------------------------------------------------
If you remember my post about sub-$100 sub-50g watches, I mentioned a handful
of them based on Miyota 2035 movements, like Bertucci A-1R and Casio
MQ-24/MQ-71. However, I also mentioned some digital ones that could serve
the same purpose while obviously offering much more functionality for the
price *and* having rather longer battery life as well. And today, I'd like
to talk about the model that usually goes under the radar but is
nevertheless significant in several ways and, as such, isn't going to leave
my collection anytime soon. In terms of price, materials, design and
functionality, it's pretty much an antipode of my ultimate recent
acquisition (Citizen PMD56-2951), but this is the case when the popular
stereotype "digital watches are bought for functionality, analogue watches
are bought to please your soul" gets busted again: this watch isn't
something to boast about, it's something to please my soul, partially
exactly because how feature-packed and retro-looking it is. Enter... Casio
DB-36.

Yes, DB-36. Not 360. I didn't need the chrome-y bling that would peel off
over time anyway because the case on the DB-360 still is plastic, I didn't
need the shiny buttons that would decrease the water resistance rating, and
I didn't need a crappy bracelet I'd have an urge to change ASAP. DB-36, on
the other hand, looks as humble as it can, almost too cheap for what it's
capable of (nowhere as cheap as the Skmei knockoffs though). By the way,
I've got rid of all the other databank-enabled Casios from my collection
(AW-80, AMW-870, G-2900F and the one I only wore once, DBC-611) for various
reasons, but this one, with its resin case, PUR strap and acrylic glass
crystal, is a survivor. I know the strap is a weak point though so I'm
probably gonna find a black 18mm mesh if I'm going to wear it more actively,
but it's no big deal if it breaks. And that's a huge upside of this watch:
even if something breaks, it's no big deal. I remember a vlogger whining
about how he broke his PMD56 (not really breaking it but ruining the WR
protective seal), telling that, like, he sold his collection to get it and
then it was broken (again, not really). Well, a $40 watch like DB-36 is much
easier to mess around with than a $600 watch like PMD56. That is, if you
ever need to.

And it turns out you normally don't: like G-2900F and other Casio watches
developed in the first half of 2000s, DB-36 (as well as 360, it's the same
module 2515 inside) is powered by a CR2025 battery that allows it to run for
over 10 years. As far as I know, the only other Databank model with the same
battery is the calculator-featuring DBC-32. All other watches in the series,
AFAIK, run on CR1616 at best. Curiously enough, old ana-digi Twincepts (ABX
series) did run on CR2016 but, of course, that 10-year lifespan wasn't
guaranteed and was close to 5 years IRL. Besides occasional battery and
strap changes, the only other piece of maintenance I can think of regarding
DB-36 would be polishing the crystal if it picks too many scratches. Which
is, again, much easier to do because it's acrylic. For what it's worth, this
watch is meant for daily-driving and not showing off, but there's even more
to that. And I'll return to this "more" after briefly going over its
functionality.

And when it comes to functionality, I'm still impressed. I mean, I hadn't
seen anything like this in any analogue, ana-digi or digital watch for <=$40
except this one. Yes, I've had an AW-80 which is close enough but it still
is more expensive. First, you've got a selection of 13 languages to display
the weekday and to determine the data entry characters. Since the module had
been developed back in 2002, I won't bash the language set too much. Then,
you've got a databank itself, which, with its 30 records, functionally
resembles the AW-80 Telememo but has a but different charset for both name
and number parts, and while the name part still is 8 characters long, the
number part is only 15 characters long, one less than in the AW-80. Then,
you've got five independent alarms (as opposed to 3 in AW-80) with the
ability to set them daily, monthly, for a specific date or for every day on
a specific month. Hourly chime and a snooze mode for the AL1 included in the
box. Then, you've got a rather typical stopwatch and a no less typical
countdown timer, both supporting full 24-hour measurement. Finally, you've
got a dual-time mode, which might *seem* like a downgrade from the AW-80's
world time mode but it's much more practical as the Reverse button in the
timekeeping mode serves as a quick-access button to the second time if you
keep it pressed for 2 seconds. Conversely, by the way, the Forward button
serves as a quick-access button to view the recently accessed databank entry.

I haven't discussed any peculiarities yet but you might have guessed where
I'm heading now that's the features are on the table. Stemming from the
Casio's lineage (no pun intended) of 1980s/1990s Databanks and having
undergone some cost-driven simplification, DB-36 is a minimum viable product
of a watch specifically designed for people WHO DON'T RELY UPON CELLPHONES
AT ALL. There, I said it. And the fact that it even can be bought new in
2024 speaks better than anything else: there still is some market for such
watches. Indeed, some people can organize their lives around desktop
computers, desktop phones (including various IP telephony), pen and paper
and wristwatches. I even think most people can do that but they just don't
want to. I can resonate with that mindset myself in some security models
when the cellphone is used in an extremely hostile environment where you
must be prepared to lose it or erase all the data from it at any moment. In
2024, a phone (as well as a paper notebook, for that matter) is much more
likely to be searched or confiscated than a cheap Casio digital watch
(unless it says F-91W, because, you know... stereotypes). If you store your
information securely enough, such a watch could contain some breadcrumbs
you, and only you, could recover it with.

Speaking of which, I'd also like to remind you about the Telememer ([1])
project. Yes, I plan to port it to a standalone CLI program too, probably in
Python as it has somewhat nice big integer support, but for now it's
implemented as a library in BigInt-only JS. The idea of Telememer is simple:
given the alphabet sets for both name and number parts, as well as their
lengths, convert arbitrary binary data into a set of records for
databank-enabled Casio watches, and, of course, be able to perform the
reverse conversion and reconstruct the source data from these records. As
the phonebooks in these watches are forcefully auto-sorted, the first
character in the name determines the index position of the record, thus
detracting from the amount of bits that can be stored this way. Yes, it
doesn't save you from the hassle of entering all the records manually and
then retyping them into the program when you need to recover the data, but
at least it allows you to store much more types of info than the plain
Telememo records do. With DB-36, however, it gets even more interesting as
different interface languages have different amount of input characters,
ranging from 46 for English to whopping 60 for French. This partially
compensates for the disadvantage of only having 15 characters available for
the numeric part and only 14 "digits" available, making one record in French
hold about |log2(14 ** 15)| + |log2(60 ** 7)| = 98 bits (considering the
first letter is reserved for the index) with the entire set holding up to
2940 bits or 367 bytes, which is 11 bytes lower than the theoretical limit
for AW-80 but still impressive.

"But wait, there's more!" If you really want to maximize your storage
capacity by fully sacrificing the convenience of data entry and having to
switch the interface language every now and then, you can still do this!
Omitting two sets of very similar letters, you have 119 distinct characters
to operate on, and the above formula will give you... well... 105 bits per
record and 3150 bits or 393 bytes per the entire set, which is more than for
AW-80. I guess that's the theoretical limit of how much you can squeeze out
of the DB-36's databank, but whether or not it's worth all the entry trouble
is up to you. For comparison, English-only entry will yield 95 bits per
record and 2850 bits/356 bytes per the entire databank. Still over 2048 bits
so think for yourselves. When porting Telememer onto Python, I'll also
prepare some character sets to make it easier for you to select which model
and language you're going to use. Not that I expect a lot of people to
actually practically use it for encoding valuable information into their
watches, but I hope it might help someone in a "life or death" situation,
and also potentially am going to use it myself. I mean, if you sacrifice the
first letter for indexing, you could even store some BIP-39 mnemonic phrases
for crypto wallets: not a great idea from the first glance, but if
absolutely necessary, why not?

Last but not least, DB-36 also functions quite well as... a watch. The screen
is pretty legible and clear, the accuracy is a bit of a lottery from item to
item but tolerable (although I haven't worn it for a prolonged period of
time yet to measure it properly), the seconds reset with the Forward button
is extremely easy without having to remove the watch from the wrist, and the
green backlight is just great. The only two visible drawbacks I could notice
are: 1) there doesn't seem to be a way to switch the backlight duration from
1.5 to 3 seconds, 2) there doesn't seem to be a way to mute the buttons,
i.e. turn off the button press confirmation beep. Curiously enough, both of
these issues had been mitigated in DBC-32, but some other trade-offs had
been made there that I'm not willing to accept. Another very important
aspect of DB-36 is that it's extremely lightweight and weighs just 27 grams
on the stock strap. I can bear 150-gram watches, I got used to 100-gram
watches, I'm pretty comfortable with 50-gram watches, but sub-30-gram ones
just disappear on the wrist until I actually need to look at them. This is,
like, the densest concentration of functionality per gram I have ever seen.
And per dollar too, by the way. I'm also going to change the 18mm strap for
a textile one from my W-800H mod (that came from a variant of AE1200WH) as
it is more durable over time and definitely is gonna make the watch look
more stylish while not increasing its weight significantly. As an
alternative, a 18mm steel mesh is being considered.

Either way, DB-36 is a quintessential digital Casio in many aspects and the
biggest "bang for a buck" when it comes to purely LCD-based models still in
production, followed only by DW-5600E and G-2900F. Despite some of its
flaws, it can be a remedy for those looking for a more disconnected
lifestyle, it's not heavy on the wrist both literally and figuratively, and
its affordability reduces the amount of headache if anything happens. Highly
recommend it, and if you can change the strap, doubly so.

--- Luxferre ---

[1]: https://gist.github.com/plugnburn/63cc2825e02311a617af55653aecef1a