On 12/8/22 12:36 PM, Charlie Gibbs wrote:
> On 2022-12-08, 26C.Z969 <26C.Z969@noaada.net> wrote:
>
>> "Something that works - quick" is usually the starting
>> point and it's usually MESSY AS HELL - on-the-fly changes
>> and squeeze-ins all over the place. This just isn't
>> good enough IMHO, un-aesthetic and you'll never figure
>> it out a year or two later.
>
> Whenever someone asks me for a "quick and dirty" fix I point out
> that "quick and dirty" is never quick, but always _very_ dirty

Oh, a note here ... You really shouldn't TELL The Bosses
that ... THEIR job is to get salable Product like YESTERDAY.
They don't UNDERSTAND programming fer shit. It's really
really like a Dilbert comic.

On 08/12/2022 08:48, The Natural Philosopher wrote:
> On 07/12/2022 22:19, Pancho wrote:
>> On 07/12/2022 09:46, The Natural Philosopher wrote:
>>> On 07/12/2022 05:14, 26C.Z969 wrote:
>>>> So, as a PERSONAL HOBBY thing, pushing Bash to its
>>>>    extremes may be a fun pursuit - but in a PRODUCTION
>>>>    ENVIRONMENT (ya know, where people PAY you to make
>>>>    software) it's a BAD THING to go too far into the
>>>>    Dark Side.
>>>
>>> Guy who really taught me how to write code that customers would pay
>>> for, said 'simple. clean, easily understood, boring workmanlike code'.
>>>
>>
>> Customers want code that works.
>>
>> Code should be simple, but it is often hard to see the simple
>> solution. If you do write a brilliant simple solution, people will
>> look at it, and say the code doesn't do very much, anyone could have
>> done it.
>>
> And you say 'next time, then hire them instead of me'
>

They never seem to care that much. They'd bump my pay, but that was it.

I have once said I couldn't develop a proposed solution, as it was too
complex, and that they should get someone else. They kept bullying me to
do it for days, I thought I was going to get fired.

In the end, they did get someone else, project failed, he was fired. I
was then given carte blanche to deliver "something".

On 09/12/2022 06:06, 26C.Z969 wrote:
> On 12/7/22 5:19 PM, Pancho wrote:
>> On 07/12/2022 09:46, The Natural Philosopher wrote:
>>> On 07/12/2022 05:14, 26C.Z969 wrote:
>>>> So, as a PERSONAL HOBBY thing, pushing Bash to its
>>>>    extremes may be a fun pursuit - but in a PRODUCTION
>>>>    ENVIRONMENT (ya know, where people PAY you to make
>>>>    software) it's a BAD THING to go too far into the
>>>>    Dark Side.
>>>
>>> Guy who really taught me how to write code that customers would pay
>>> for, said 'simple. clean, easily understood, boring workmanlike code'.
>>>
>>
>> Customers want code that works.
>
>   And the companies providing that code want
>   it to be MAINTAINABLE/EXPANDABLE.
>
>   Gobbledegoop Bash/C/Python/Perl/Java/Whatever
>   does NOT fit that criteria, no matter how 'clever'.
>

In my experience, not so much, mainly they just want it to work, now!

>> Code should be simple, but it is often hard to see the simple
>> solution. If you do write a brilliant simple solution, people will
>> look at it, and say the code doesn't do very much, anyone could have
>> done it.
>
>   Ha !
>
>   Oh, and "brilliant solutions" can often be expressed
>   simply and clearly in code.

Any system will have a minimum level of simplicity. I would argue one of
the metrics of code quality is maximising simplicity. It's not just me
that argues this, Albert Einstein once said “Everything should be made
as simple as possible, but not simpler.”

In the four decades I have worked in software development, I have seen
massive advances simplifying code, to achieve similar results. Overall
complexity has increased, but that is because software now does a lot more.

But at my level I have often seen overly complex systems lauded as
brilliant, worthy of respect and funding, whereas simple systems are
dismissed as trivial. If a project costs a lot, managers big up its
value, to justify the cost.

In one instance, I can remember being rejected at an interview as a
developer for a project. 18 months later, after this project had failed.
I was then employed to develop a new project to achieve the same goal. I
sat at the former desk of the failed project manager, who rejected me,
and who had just been fired. Six months later, with my solution a huge
success, another manager who had also been in my initial interview and
rejected me, took over the project, in order to make the code "a
corporate core technology". His comment was that it didn't really do
anything. No mention of the previous failure.

Politics is what counts, very rarely do they care about having good
developers. In my career, a few periods when they were desperate, but
that is it.

On 12/9/22 5:25 AM, Pancho wrote:
> On 09/12/2022 06:06, 26C.Z969 wrote:
>> On 12/7/22 5:19 PM, Pancho wrote:
>>> On 07/12/2022 09:46, The Natural Philosopher wrote:
>>>> On 07/12/2022 05:14, 26C.Z969 wrote:
>>>>> So, as a PERSONAL HOBBY thing, pushing Bash to its
>>>>>    extremes may be a fun pursuit - but in a PRODUCTION
>>>>>    ENVIRONMENT (ya know, where people PAY you to make
>>>>>    software) it's a BAD THING to go too far into the
>>>>>    Dark Side.
>>>>
>>>> Guy who really taught me how to write code that customers would pay
>>>> for, said 'simple. clean, easily understood, boring workmanlike code'.
>>>>
>>>
>>> Customers want code that works.
>>
>>    And the companies providing that code want
>>    it to be MAINTAINABLE/EXPANDABLE.
>>
>>    Gobbledegoop Bash/C/Python/Perl/Java/Whatever
>>    does NOT fit that criteria, no matter how 'clever'.
>>
>
> In my experience, not so much, mainly they just want it to work, now!

Those are the companies that go out-of-biz TOMORROW.

Don't fall for the pension-plan/stock-options ploy.

>>> Code should be simple, but it is often hard to see the simple
>>> solution. If you do write a brilliant simple solution, people will
>>> look at it, and say the code doesn't do very much, anyone could have
>>> done it.
>>
>>    Ha !
>>
>>    Oh, and "brilliant solutions" can often be expressed
>>    simply and clearly in code.
>
> Any system will have a minimum level of simplicity. I would argue one of
> the metrics of code quality is maximising simplicity. It's not just me
> that argues this, Albert Einstein once said “Everything should be made
> as simple as possible, but not simpler.”

So far agreed ... but some here (you?) seem to be
glorifying gobbldegoop code.

> In the four decades I have worked in software development, I have seen
> massive advances simplifying code, to achieve similar results. Overall
> complexity has increased, but that is because software now does a lot more.

Big changes since the 8008 chip (which I did write some
code for) ....

> But at my level I have often seen overly complex systems lauded as
> brilliant, worthy of respect and funding, whereas simple systems are
> dismissed as trivial. If a project costs a lot, managers big up its
> value, to justify the cost.

That CAN happen. However said pointy-haired bosses
usually bankrupt the company. Forget the vacation
plan too ......

When you find yourself in a situation where close
management simply does not understand code, TIME
TO SLIP OUT THE BACK DOOR.

> In one instance, I can remember being rejected at an interview as a
> developer for a project. 18 months later, after this project had failed.
> I was then employed to develop a new project to achieve the same goal. I
> sat at the former desk of the failed project manager, who rejected me,
> and who had just been fired. Six months later, with my solution a huge
> success, another manager who had also been in my initial interview and
> rejected me, took over the project, in order to make the code "a
> corporate core technology". His comment was that it didn't really do
> anything. No mention of the previous failure.

Good "moral revenge"/karma story. Lucky the corp survived
to PUT you at that desk. Lots and lots of "software"
companies are undercapitalized and lack product diversity and
so the first failure is the LAST failure.

> Politics is what counts, very rarely do they care about having good
> developers. In my career, a few periods when they were desperate, but
> that is it.

As the chain of command gets longer, the instances of utter,
damaging, maybe fatal, STUPIDITY rises exponentially. Those
In Charge will NOT understand their products or its production
in the least.

The BEST solution is to migrate to the smaller outfits, places
where your expertise is appreciated, where the bosses are likely
to take good advice. The pay may not be quite as great but the
satisfaction/security index more than makes up for that.

But STILL don't write gobbledgoop code ....

And, on theme, pushing Bash scripts to their ultimate
incomprehensible limits IS gobbledegoop. I'd say ONE
level above DOS batch files is as far as you should go
before switching to a higher-level language. Then
Those Who Come After will PRAISE you and the company
will survive to pay yer pension.

. .

I'm near retirement at this point. This presents a
problem for management. I'm THE "guru", the only one
who gets the whole scheme, knows how to keep it all
going and safe and how to improve all that. There is
NOT "another me" in the pipeline and won't be - they
were lucky to benefit from my multiple software/hardware/
systems skills for decades (hey, if The Device doesn't
exist - MAKE the damned thing ! Chips & solder & assembler !)

They are too small to attract such talent now. No kiddies
with the skills would hang around for more than a year
or so but (foolishly) migrate to the Big Boyz in hopes
of Big $$$. So, little "continuity".

The fill-ins are a level or two below me - DO have
some code/OS skills but aren't gonna be MAKING anything
of consequence that's "new/useful". I have, per PWE
(even though I'm 'ethical' but not religious), been
setting them up to "coast" for a year or two.
After that .....

So ... they're doing stuff that's not entirely wise.
Spending large quantities of money trying to externalize
services, standardize (alas with horrid MS products)
under the theory that SOME commercial IT provider will
at least keep the mail/office/net stuff going - not
understanding that there's a LOT just below the surface
that really NEEDS to keep running also (or grasping
that MS is just the visible tip of the proverbial
iceberg and that everything else runs on Linux/Unix
right down to the fun little gadgets they depend on).

Alas the 'providers' we've had experience with have
NOT been very encouraging in these aspects. We're Just
Another Small-ish Account. My GUESS is that they'll
perish from glitches, malware and general MS idiocy
within a couple of years. Didn't want it to be that
way but, well, currently, people DO get old.

So now I'm between the rock and hard place. I don't
want 30+ years of labor go to nothing, but it looks
like it WILL go that way. I don't want the org and
people to suffer, but it looks like it WILL go
that way. For what they WANT to do/be, those external
providers AIN'T gonna do it for them. Related orgs
also had olde-dayz "gurus" and now they're pretty
much all gone too.

Words of wisdom would be appreciated. How do we
attract/HOLD latter-day "gurus" ??? Maybe ply
the Asperger/Autistic ranks ??? (I admit to
being lightly Aspie, made me a good compuGeek,
though there was no such term when I was in
the schools :-)

On 10/12/2022 07:48, 26C.Z969 wrote:
> Words of wisdom would be appreciated. How do we
>   attract/HOLD latter-day "gurus" ??? Maybe ply
>   the Asperger/Autistic ranks ??? (I admit to
>   being lightly Aspie, made me a good compuGeek,
>   though there was no such term when I was in
>   the schools

I am afraid the simple answer is that software is going to the dogs.
Originally people paid you to make stuff work. Now they pay you to fit
some corporate ideal in terms of development methodology, moral
ideology, and political alignment.

I used to like playing Bridge, There is a good online site, but its full
of Americans who have been taught to play the game in one way only.

I opened, my partner responded and I simply said 'six spades' It was
obvious what he had, given what I held, there was no need to go through
the process of establishing it and revealing any weakness to the opposition.

We made the contract but the table's host removed me for 'not playing
properly'!

I am afraid that the Not So Smart give more value to doing things the
way they were taught than achieving any actual results.

Yesterday some pointy-head decreed that the London Fire Service were
misogynistic and institutionally racist....

....No mention was made of their ability, or otherwise, to *put out fires*.

One is tempted to recommend the 'Peter Principle', and, more chillingly
Joseph Tainter's 'Collapse of complex societies'.

--
WOKE is an acronym... Without Originality, Knowledge or Education.

On 10/12/2022 07:48, 26C.Z969 wrote:

>   Good "moral revenge"/karma story. Lucky the corp survived
>   to PUT you at that desk. Lots and lots of "software"
>   companies are undercapitalized and lack product diversity and
>   so the first failure is the LAST failure.
>

Credit Suisse, just about survived. :-)

On 12/10/22 9:36 AM, Pancho wrote:
> On 10/12/2022 07:48, 26C.Z969 wrote:
>
>>    Good "moral revenge"/karma story. Lucky the corp survived
>>    to PUT you at that desk. Lots and lots of "software"
>>    companies are undercapitalized and lack product diversity and
>>    so the first failure is the LAST failure.
>>
>
> Credit Suisse, just about survived. :-)

Let's wait to see how much they were invested
in FTX Crypto :-)

In any case, CS would be considered a "diverse"
company - a broad line of profitable products
and investments. A few disasters here and there
have little impact.

On 12/10/22 5:09 AM, The Natural Philosopher wrote:
> On 10/12/2022 07:48, 26C.Z969 wrote:
>> Words of wisdom would be appreciated. How do we
>>    attract/HOLD latter-day "gurus" ??? Maybe ply
>>    the Asperger/Autistic ranks ??? (I admit to
>>    being lightly Aspie, made me a good compuGeek,
>>    though there was no such term when I was in
>>    the schools
>
> I am afraid the simple answer is that software is going to the dogs.
> Originally people paid you to make stuff work. Now they pay you to fit
> some corporate ideal in terms of development methodology,  moral
> ideology, and political alignment.
>
> I used to like playing Bridge, There is a good online site, but its full
> of Americans who have been taught to play the game in one way only.
>
> I opened, my partner responded and I simply said 'six spades' It was
> obvious what he had, given what I held, there was no need to go through
> the process of establishing it and revealing any weakness to the
> opposition.
>
> We made the contract but the table's host removed me for 'not playing
> properly'!
>
> I am afraid that the Not So Smart give more value to doing things the
> way they were taught than achieving any actual results.
>
> Yesterday some pointy-head decreed that the London Fire Service were
> misogynistic and institutionally racist....
>
> ...No mention was made of their ability, or otherwise, to *put out fires*.

A useful skill - but that doesn't mean they ought
to be total barbarians. Doesn't mean they should
be total idiots either. Embrace those who CAN, and
reject those who CANNOT without apologies to the
political fads of the moment.

> One is tempted to recommend the 'Peter Principle',  and, more chillingly
> Joseph Tainter's 'Collapse of complex societies'.

I sense that western civ, 'democracy', is coming up to
a BIG test very soon. Dictatorship has been the global
norm for upteen thousand years - 'free'/'democratic'
states are just a recent flash in the pan by comparison.
In some cases, their strengths are also their worst
weaknesses. Bickering and ideological clashes replace
practical results.

On 11/12/2022 04:06, 26C.Z969 wrote:
> On 12/10/22 9:36 AM, Pancho wrote:
>> On 10/12/2022 07:48, 26C.Z969 wrote:
>>
>>>    Good "moral revenge"/karma story. Lucky the corp survived
>>>    to PUT you at that desk. Lots and lots of "software"
>>>    companies are undercapitalized and lack product diversity and
>>>    so the first failure is the LAST failure.
>>>
>>
>> Credit Suisse, just about survived. :-)
>
>   Let's wait to see how much they were invested
>   in FTX Crypto  :-)
>

I very much doubt it is significant. The proof of work/blockchain always
seemed a very bad idea to me. If you have reliable/reputable servers,
and we do, there are so much more efficient ways to distribute a ledger.

I see CS have current difficulties, but I've no idea what they are. It
is a long time since I took any interest in that kind of thing.

>   In any case, CS would be considered a "diverse"
>   company - a broad line of profitable products
>   and investments. A few disasters here and there
>   have little impact.
>

I was just making the point that the antics of a few programmers are
small potatoes to such a company. The failed project would not have cost
more than 1 to 3 million GBP. The solution I delivered significantly
less. I was a cheap RAD developer. Knock something up, if there was a
need for more it was handed over to more strategic developers, who
adapted or rewrote it. At the time it pissed me off that others were
given better resources and timescales than me, but in hindsight I see it
is what I was good at.

When I worry about the impact of the work I did, it is my “successful”
projects that worry me, not the failures. Worried others, too. :-)

26C.Z969 <26C.Z969@noaada.net> wrote:
> On 12/5/22 4:59 PM, Computer Nerd Kev wrote:
>
>> I could have posted a rant here a few years ago complaining about
>> the compatibility issues with Python, and how everyone who's
>> switched to it should "consider" going back to Bash, for example.
>> But I didn't, because it's just my preference against theirs, and I
>> just have to put up with less stuff published online being done
>> with Bash scripts, and working in ten years time.
>
> The "compatibility issues" with Python are largely
> illusory. The main thing was P2 to P3. We'll have
> to wait to see what happens with P4. Even with
> P2->P3 I'd say there was 95% backwards compatibility,
> esp in the "understandability" areas.

I saw this today. So if it's all "illusory", well I'm not the only
person having this dream.

Python version upgrades and deprecations
By Chris Siebenmann, December 8, 2022
- https://utcc.utoronto.ca/~cks/space/blog/python/PythonUpgradesAndDeprectation

"Recently I read Itamar Turner-Trauring's It's time to stop using
Python 3.7 (via). On the one hand, this is pragmatic advice,
because as the article mentions Python 3.7 is reaching its end of
life as of June 2023. On the other hand it gives me feelings, and
one of the feelings is that the Python developers are not making
upgrades any easier by slowly deprecating various standard library
modules. Some of these modules are basically obsolete now, but some
are not and have no straightforward replacement, such as the cgi
module.

The Python developers can do whatever they want to do (that's the
power of open source), and they clearly want to move Python forward
(as they see it) by cleaning up the standard library. But this
means that they are perfectly willing to break backward
compatibility in Python 3, at least for the standard library.

One of the things that make upgrading versions of anything easy is
if the new version is a drop in replacement for the old one.
Deprecating and eventually removing things in new versions means
that new versions are not drop in replacements, which means that it
makes upgrading harder. When upgrading is harder, (more) people put
it off or don't do it. This happens regardless of what the software
authors like or think, because people are people.

I doubt this is a direct factor in people still using Python 3.7.
But I can't help but think that the Python developers' general
attitude toward backward compatibility doesn't help." ...

--
__ __
#_ < |\| |< _#

On 12/11/22 5:49 AM, Pancho wrote:
> On 11/12/2022 04:06, 26C.Z969 wrote:
>> On 12/10/22 9:36 AM, Pancho wrote:
>>> On 10/12/2022 07:48, 26C.Z969 wrote:
>>>
>>>>    Good "moral revenge"/karma story. Lucky the corp survived
>>>>    to PUT you at that desk. Lots and lots of "software"
>>>>    companies are undercapitalized and lack product diversity and
>>>>    so the first failure is the LAST failure.
>>>>
>>>
>>> Credit Suisse, just about survived. :-)
>>
>>    Let's wait to see how much they were invested
>>    in FTX Crypto  :-)
>>
>
> I very much doubt it is significant. The proof of work/blockchain always
> seemed a very bad idea to me. If you have reliable/reputable servers,
> and we do, there are so much more efficient ways to distribute a ledger.

But "blockchain" sounds so COOL to the managerial
set - something they can brag about ! :-)

Pointless/useless ... irrelevant so long as there
are short-term profits and promotions.

> I see CS have current difficulties, but I've no idea what they are. It
> is a long time since I took any interest in that kind of thing.

Well, almost EVERYWHERE has "current difficulties". Combine
the usual managerial idiocy with years of destructive Covid
over-response and, well .....

>>    In any case, CS would be considered a "diverse"
>>    company - a broad line of profitable products
>>    and investments. A few disasters here and there
>>    have little impact.
>>
>
> I was just making the point that the antics of a few programmers are
> small potatoes to such a company.

As said, they're "big" and diverse. This is protective.

But even THEN ... wise programmers would seem a better
investment than a gaggle of idiot programmers. Banks
ought to understand "investments".

> The failed project would not have cost
> more than 1 to 3 million GBP. The solution I delivered significantly
> less. I was a cheap RAD developer. Knock something up, if there was a
> need for more it was handed over to more strategic developers, who
> adapted or rewrote it. At the time it pissed me off that others were
> given better resources and timescales than me, but in hindsight I see it
> is what I was good at.

You were "cheap labor" - so you got the corresponding
salary/recognition ....

I like RAD development/environments myself. My fav is Lazarus/FPC.
If you need a passable GUI by friday, that's the way to go. All-
purpose script languages like Python are also good for whipping
up something usable quick. Plain old HTML/PHP can deliver functional
browser interfaces pretty quick as well (never liked javascript, but
a sprinkling here and there CAN add a little eye-candy).

Seems like whatever I write it in FIRST, a week or two later
I re-write it - "better" - in something else. This week I'm
making a functional clone of a recent FPC app in Python3 because
the likely Next Guy to use/tweak it is mostly only good at
Python (and Winders, so I've gotta do it cross-platform). In this
case it's "better" because it has a longer future, not that Python
is a "better" language than Pascal( ok, ok ... DO like the string-
slicing and EZ lists and they have a library for EVERYTHING ...)

I do like Pascal ... you can do almost anything you can do
in 'C' but it's more elegant and for sure more self-documenting.
Got into that with the MS/IBM multi-pass Pascal compiler for
early DOS in the pink ring-binders. STILL sorta have it too ...
in a DOS VM ... just for fun. However Turbo made a gigantic
diff in the utility of Pascal and Delphi provided the
super-tweakable slap-it-up GUI capability (until it priced
itself out of any sensible market).

> When I worry about the impact of the work I did, it is my “successful”
> projects that worry me, not the failures. Worried others, too. :-)

Uh oh ... what WERE you successful at ... something wicked to
exploit the masses ? :-)

On 12/12/22 9:09 PM, Computer Nerd Kev wrote:
> 26C.Z969 <26C.Z969@noaada.net> wrote:
>> On 12/5/22 4:59 PM, Computer Nerd Kev wrote:
>>
>>> I could have posted a rant here a few years ago complaining about
>>> the compatibility issues with Python, and how everyone who's
>>> switched to it should "consider" going back to Bash, for example.
>>> But I didn't, because it's just my preference against theirs, and I
>>> just have to put up with less stuff published online being done
>>> with Bash scripts, and working in ten years time.
>>
>> The "compatibility issues" with Python are largely
>> illusory. The main thing was P2 to P3. We'll have
>> to wait to see what happens with P4. Even with
>> P2->P3 I'd say there was 95% backwards compatibility,
>> esp in the "understandability" areas.
>
> I saw this today. So if it's all "illusory", well I'm not the only
> person having this dream.

You expect everything to stay exactly the same forever ???

Better go back to Bash 0.1 then - not use any of those
newfangled features ! Better yet, stick with CP/M-80
or System-V ... otherwise Those Bastards might CHANGE
something !

> Python version upgrades and deprecations
> By Chris Siebenmann, December 8, 2022
> - https://utcc.utoronto.ca/~cks/space/blog/python/PythonUpgradesAndDeprectation
>
> "Recently I read Itamar Turner-Trauring's It's time to stop using
> Python 3.7 (via). On the one hand, this is pragmatic advice,
> because as the article mentions Python 3.7 is reaching its end of
> life as of June 2023. On the other hand it gives me feelings, and
> one of the feelings is that the Python developers are not making
> upgrades any easier by slowly deprecating various standard library
> modules. Some of these modules are basically obsolete now, but some
> are not and have no straightforward replacement, such as the cgi
> module.

IMHO if P3.9 is current, you should write for P3.5 in order to
cover different sub-versions found on various systems. The
newest tricks are also the most likely to 'evolve' in ways
that cause issues as the "oops ! we kinda hurried THAT out"
realizations appear.

I've been writing a lot of Python this week - and it ALL works
as advertised ... and my older code STILL works.

Dunno WHAT you're doing that's SO unstable. Suggest doing it
some other way.

I'm re-writing a large-ish P3.2 script that's been in use to
this day. It does OS-level stuff, subprocesses and threads,
and STILL WORKS. The only reason I'm re-writing it is because
there'd been too much feature-creep and the logic was getting
too hard to see. The new one is much neater and structurally
incorporates the most important creepy features instead
of them being just jammed in there with a bunch of flags
and such required to make 'em work. I like neat code you
can still understand five or ten years later. Also wrote a
Pascal version for Linux - the Python version is meant to
be cross-platform (and Those Who Will Follow are a lot
more into Python than Pascal or 'C' ... Python is to
them what BASIC was to the previous gen or COBOL and FORTRAN
were to the even older geeks - it's what they learned in
school, super-familiar do-it-alls).

> The Python developers can do whatever they want to do (that's the
> power of open source), and they clearly want to move Python forward
> (as they see it) by cleaning up the standard library. But this
> means that they are perfectly willing to break backward
> compatibility in Python 3, at least for the standard library.

I think they break it when they HAVE to, not for funzies.

P3 was a significant internal paradigm shift after LONG
consideration and debate. I don't think P4 will be as
drastic.

Dunno what's up for the next gen ... likely Gen-5+ 'AI'-ware
where you don't so much "program" as describe the problem and
let the 'compiler' assemble the working code from a bunch of
little standard building blocks. SOMEBODY will still have to
write/maintain the underlying system of course, probably
still mostly 'C', for awhile until the 'AI' system is writ
in and by the 'AI' system. There's more than one language
that came to be "written in itself" ... sometimes including
the underlying OS too. I guess the "most famous" example
was Wirth's Lillith box

> One of the things that make upgrading versions of anything easy is
> if the new version is a drop in replacement for the old one.
> Deprecating and eventually removing things in new versions means
> that new versions are not drop in replacements, which means that it
> makes upgrading harder. When upgrading is harder, (more) people put
> it off or don't do it. This happens regardless of what the software
> authors like or think, because people are people.
>
> I doubt this is a direct factor in people still using Python 3.7.
> But I can't help but think that the Python developers' general
> attitude toward backward compatibility doesn't help." ...

Nope, I'm still gonna use P3.7 and beyond. Great language
for whipping-up stuff in a hurry and a thoughtful, somewhat
conservative, approach minimizes the depreciation issues.
And for working a lot on STRINGS in a quick comprehensible
fashion there's really nothing to beat it. I think they
stole string-slicing from FORTRAN though .....

26C.Z969 <26C.Z969@noaada.net> wrote:
> On 12/12/22 9:09 PM, Computer Nerd Kev wrote:
>> 26C.Z969 <26C.Z969@noaada.net> wrote:
>>> On 12/5/22 4:59 PM, Computer Nerd Kev wrote:
>>>
>>>> I could have posted a rant here a few years ago complaining about
>>>> the compatibility issues with Python, and how everyone who's
>>>> switched to it should "consider" going back to Bash, for example.
>>>> But I didn't, because it's just my preference against theirs, and I
>>>> just have to put up with less stuff published online being done
>>>> with Bash scripts, and working in ten years time.
>>>
>>> The "compatibility issues" with Python are largely
>>> illusory. The main thing was P2 to P3. We'll have
>>> to wait to see what happens with P4. Even with
>>> P2->P3 I'd say there was 95% backwards compatibility,
>>> esp in the "understandability" areas.
>>
>> I saw this today. So if it's all "illusory", well I'm not the only
>> person having this dream.
>
> You expect everything to stay exactly the same forever ???

No, I want it to stay as backwards compatible as possible. Bash is
good at that, Python isn't, so I prefer Bash to Python.

Adopting a language that after decades still isn't interested in
backwards compatibility is something to be done out of brutal
necessity, not a vague idea of elegance. I get it that you really
don't like Bash as a scripting language, but this is a personal
opinion of yours, other people may be more troubled by Python's
compatibility issues.

> Better go back to Bash 0.1 then - not use any of those
> newfangled features !

The point is that I can update Bash and not worry about lots of
changed/depreciated stuff breaking existing scripts. It also avoids
trouble with scripts found on the web failing because they're a
few years old.

> Better yet, stick with CP/M-80
> or System-V ... otherwise Those Bastards might CHANGE
> something !

Well alright, I am posting from a PC running Linux kernel 2.4, so
you're not all that far off.

--
__ __
#_ < |\| |< _#

On 13/12/2022 06:12, 26C.Z969 wrote:

>> I was just making the point that the antics of a few programmers are
>> small potatoes to such a company.
>
>   As said, they're "big" and diverse. This is protective.
>
>   But even THEN ... wise programmers would seem a better
>   investment than a gaggle of idiot programmers. Banks
>   ought to understand "investments".
>

During the time I worked there, the company was actually a lot less
diverse than you might imagine, Corporate P&L was often dominated by
very small areas. The point I was making was that software development
costs didn't dominate P&L.

The company was generally technically more competent than most. I figure
my benefit was that I was clever enough to know how to deliver a
solution, without being so clever I was overconfident. My work was
pragmatic and constrained by fear of failure.

The Bank staff did understand investments, it was the
government/regulators that did not.

>> The failed project would not have cost more than 1 to 3 million GBP.
>> The solution I delivered significantly less. I was a cheap RAD
>> developer. Knock something up, if there was a need for more it was
>> handed over to more strategic developers, who adapted or rewrote it.
>> At the time it pissed me off that others were given better resources
>> and timescales than me, but in hindsight I see it is what I was good at.
>
>   You were "cheap labor" - so you got the corresponding
>   salary/recognition ....
>

I just meant RAD solutions were cheap compared to strategic solutions.

>   I like RAD development/environments myself. My fav is Lazarus/FPC.
>   If you need a passable GUI by friday, that's the way to go. All-
>   purpose script languages like Python are also good for whipping
>   up something usable quick. Plain old HTML/PHP can deliver functional
>   browser interfaces pretty quick as well (never liked javascript, but
>   a sprinkling here and there CAN add a little eye-candy).
>

I liked RAD because it put me next to the users and minimised
intermediate politics. There was more pleasure, for me, to see "a plan
come together", than there is to develop small components/aspects of a
large system. I also tend to perform better with simplicity than I do
with complexity, relative to other people. e.g. I'm relatively better at
pure maths than at physics.

>   Seems like whatever I write it in FIRST, a week or two later
>   I re-write it - "better" - in something else. This week I'm
>   making a functional clone of a recent FPC app in Python3 because
>   the likely Next Guy to use/tweak it is mostly only good at
>   Python (and Winders, so I've gotta do it cross-platform). In this
>   case it's "better" because it has a longer future, not that Python
>   is a "better" language than Pascal( ok, ok ... DO like the string-
>   slicing and EZ lists and they have a library for EVERYTHING ...)
>
>   I do like Pascal ... you can do almost anything you can do
>   in 'C' but it's more elegant and for sure more self-documenting.
>   Got into that with the MS/IBM multi-pass Pascal compiler for
>   early DOS in the pink ring-binders. STILL sorta have it too ...
>   in a DOS VM ... just for fun. However Turbo made a gigantic
>   diff in the utility of Pascal and Delphi provided the
>   super-tweakable slap-it-up GUI capability (until it priced
>   itself out of any sensible market).
>
>
>> When I worry about the impact of the work I did, it is my “successful”
>> projects that worry me, not the failures. Worried others, too. :-)
>
>   Uh oh ... what WERE you successful at ... something wicked to
>   exploit the masses ?  :-)

Some of my work may have played a small part in fucking the world, but
in general I just meant the moment code is successful, more people look
at it. The "Roving eye of Sauron" looks at what I have been up to, looks
at my development compromises, my mistakes, "where the bodies are
buried" :-).

On 12/14/22 4:39 PM, Computer Nerd Kev wrote:
> 26C.Z969 <26C.Z969@noaada.net> wrote:
>> On 12/12/22 9:09 PM, Computer Nerd Kev wrote:
>>> 26C.Z969 <26C.Z969@noaada.net> wrote:
>>>> On 12/5/22 4:59 PM, Computer Nerd Kev wrote:
>>>>
>>>>> I could have posted a rant here a few years ago complaining about
>>>>> the compatibility issues with Python, and how everyone who's
>>>>> switched to it should "consider" going back to Bash, for example.
>>>>> But I didn't, because it's just my preference against theirs, and I
>>>>> just have to put up with less stuff published online being done
>>>>> with Bash scripts, and working in ten years time.
>>>>
>>>> The "compatibility issues" with Python are largely
>>>> illusory. The main thing was P2 to P3. We'll have
>>>> to wait to see what happens with P4. Even with
>>>> P2->P3 I'd say there was 95% backwards compatibility,
>>>> esp in the "understandability" areas.
>>>
>>> I saw this today. So if it's all "illusory", well I'm not the only
>>> person having this dream.
>>
>> You expect everything to stay exactly the same forever ???
>
> No, I want it to stay as backwards compatible as possible. Bash is
> good at that, Python isn't, so I prefer Bash to Python.
>
> Adopting a language that after decades still isn't interested in
> backwards compatibility is something to be done out of brutal
> necessity, not a vague idea of elegance. I get it that you really
> don't like Bash as a scripting language, but this is a personal
> opinion of yours, other people may be more troubled by Python's
> compatibility issues.
>
>> Better go back to Bash 0.1 then - not use any of those
>> newfangled features !
>
> The point is that I can update Bash and not worry about lots of
> changed/depreciated stuff breaking existing scripts. It also avoids
> trouble with scripts found on the web failing because they're a
> few years old.

I still can't figure out what you're doing that
gets depreciated so fast ......

As said, I have 8-10 year old scripts from the
dawn of P3 that still work just fine - fairly
large/complex ones.

Also as said, there's a POINT where trying to
push Bash (or any of the other shell scripts)
too far results in gobblegoop NOBODY is going
to appreciate (or wanna pay for).

"Backwards compatibility" is only just SO good
an excuse. There are a relative gazillion people
who can easily, quickly, fix-up the odd depreciation
issues in Python. It's the most popular computer
language in the world by far and is going to STAY
that way for decades. Endless talent, much a lot
cheaper than ye or me. Think about it.

There are perfectly good alternatives to Python
of course, assuming you're afraid of snakes.
There's always Perl (hate it but it CAN do a
lot) even Java (hate it a lot, even though it
CAN do a lot (in an excessively wordy and
weird fashion)).

>> Better yet, stick with CP/M-80
>> or System-V ... otherwise Those Bastards might CHANGE
>> something !
>
> Well alright, I am posting from a PC running Linux kernel 2.4, so
> you're not all that far off.

A bit dated ... but still not TOO bad. THE issue, at least
for commercial/govt clients, is whether they're keeping up
with the security issues. Those alone may 'depreciate'
certain solutions. Gotta keep up with Putin/Xi/Kim.

I have taken a shot at the BSDs ... theoretically "more
stable/secure". Alas even with some relatively simple
things I can't seem to get there from here. So far as
I can tell they're STILL stuck at SMB1 or SMB2 and I
want/need the more advanced security/speed. A LOT of
what's in the BSDs seems a decade (or two) behind
the curve. Try 'em, I'm sure you'd love 'em :-)

I cut my proverbial teeth on a PDP-11 with that
newfangled 'C' and a little FORTRAN (the OLD
FORTRAN where everything had to be in the exact
right columns on the punch-cards). The situation
HAS improved. Some of that improvement involved
a loss of some backwards compatibility. I roll
with the punches because if I don't someone
younger/cheaper/faster damned well WILL. I just
need to scoot to 65 or 66 and then they can
have it ..... (I foresee disaster) .........

Still have my K&R book on 'C' ... and DO look
at it from time to time. The modern tweaks can
be great, but they're also the most likely to
mutate. The olde straight to-the-point approach
can be superior sometimes (and compiles JUST
as fast and tight because the 'New Ways' are
just wallpaper over the Olde Ways. I would
suggest "Algorithms in C" by Robert Sedgewick
though as a K&R adjunct.

Finally, Python is just a double-layer of wallpaper
over 'C' - you can just FEEL it sometimes ... linked
lists of linked lists and such .......... wrote
some of those kinds of things in modern FORTRAN and
ADA, just for fun (hate ADA too - needs a giant dose
of laxative and sphincter relaxant, I don't think
the actual Ada would have liked it). Whatever became
of PL/I, one of the great "and the kitchen sink"
languages ???? :-)

On 15/12/2022 06:43, 26C.Z969 wrote:
>> The point is that I can update Bash and not worry about lots of
>> changed/depreciated stuff breaking existing scripts. It also avoids
>> trouble with scripts found on the web failing because they're a
>> few years old.
>
>   I still can't figure out what you're doing that
>   gets depreciated so fast ......
>
Especially as you never paid for it in the first place...

--
It’s easier to fool people than to convince them that they have been fooled.
Mark Twain

On 14/12/2022 21:39, Computer Nerd Kev wrote:
> I am posting from a PC running Linux kernel 2.4, so
> you're not all that far off.

If it works for you...why not? It didn't work for me back in the day,
which is why I run recent kernels

--
It’s easier to fool people than to convince them that they have been fooled.
Mark Twain

On 12/15/22 5:10 AM, The Natural Philosopher wrote:
> On 15/12/2022 06:43, 26C.Z969 wrote:
>>> The point is that I can update Bash and not worry about lots of
>>> changed/depreciated stuff breaking existing scripts. It also avoids
>>> trouble with scripts found on the web failing because they're a
>>> few years old.
>>
>>    I still can't figure out what you're doing that
>>    gets depreciated so fast ......
>>
> Especially as you never paid for it in the first place...

Well ... I'll give some credit to "hobbyists", but
yea, when you do it for a LIVING you can't always
be such a "purist". The Boss not only wants results
but results The Other Guys can DO stuff with.

On 12/15/22 5:12 AM, The Natural Philosopher wrote:
> On 14/12/2022 21:39, Computer Nerd Kev wrote:
>> I am posting from a PC running Linux kernel 2.4, so
>> you're not all that far off.
>
> If it works for you...why not? It didn't work for me back in the day,
> which is why I run recent kernels

He's into "purism" ... which is good in some ways
and not-so-good in others.

I've no doubt he can coerce Bash into doing all
kinds of weird things. However while that's fine
for a purist "civvie" it is NOT good for anybody
who expects a paycheck for programming (and there
are endless newbies/pretenders out there who'll
promise better for less).

If he likes it Olde Style he really ought to try
some of the BSDs. They progress REALLY slow :-)

On 16/12/2022 04:26, 26C.Z969 wrote:
> On 12/15/22 5:10 AM, The Natural Philosopher wrote:
>> On 15/12/2022 06:43, 26C.Z969 wrote:
>>>> The point is that I can update Bash and not worry about lots of
>>>> changed/depreciated stuff breaking existing scripts. It also avoids
>>>> trouble with scripts found on the web failing because they're a
>>>> few years old.
>>>
>>>    I still can't figure out what you're doing that
>>>    gets depreciated so fast ......
>>>
>> Especially as you never paid for it in the first place...
>
>   Well ... I'll give some credit to "hobbyists", but
>   yea, when you do it for a LIVING you can't always
>   be such a "purist". The Boss not only wants results
>   but results The Other Guys can DO stuff with.

....whoosh....

--
It is the folly of too many to mistake the echo of a London coffee-house
for the voice of the kingdom.

Jonathan Swift

On 12/16/22 4:24 AM, The Natural Philosopher wrote:
> On 16/12/2022 04:26, 26C.Z969 wrote:
>> On 12/15/22 5:10 AM, The Natural Philosopher wrote:
>>> On 15/12/2022 06:43, 26C.Z969 wrote:
>>>>> The point is that I can update Bash and not worry about lots of
>>>>> changed/depreciated stuff breaking existing scripts. It also avoids
>>>>> trouble with scripts found on the web failing because they're a
>>>>> few years old.
>>>>
>>>>    I still can't figure out what you're doing that
>>>>    gets depreciated so fast ......
>>>>
>>> Especially as you never paid for it in the first place...
>>
>>    Well ... I'll give some credit to "hobbyists", but
>>    yea, when you do it for a LIVING you can't always
>>    be such a "purist". The Boss not only wants results
>>    but results The Other Guys can DO stuff with.
>
> ...whoosh....

Perhaps ...

No, he never PAID for Python. Takes balls to bitch
about a free product :-)

But I'm still confused what HE'S doing that seems
to get horribly depreciated SO quickly .....

In any case, perhaps HE can afford to be a "purist",
but we who DO earn money from the craft can't be
so picky all the time. Sometimes you DO have to
work and play well with others.

On 17/12/2022 06:29, 26C.Z969 wrote:
> On 12/16/22 4:24 AM, The Natural Philosopher wrote:
>> On 16/12/2022 04:26, 26C.Z969 wrote:
>>> On 12/15/22 5:10 AM, The Natural Philosopher wrote:
>>>> On 15/12/2022 06:43, 26C.Z969 wrote:
>>>>>> The point is that I can update Bash and not worry about lots of
>>>>>> changed/depreciated stuff breaking existing scripts. It also avoids
>>>>>> trouble with scripts found on the web failing because they're a
>>>>>> few years old.
>>>>>
>>>>>    I still can't figure out what you're doing that
>>>>>    gets depreciated so fast ......
>>>>>
>>>> Especially as you never paid for it in the first place...
>>>
>>>    Well ... I'll give some credit to "hobbyists", but
>>>    yea, when you do it for a LIVING you can't always
>>>    be such a "purist". The Boss not only wants results
>>>    but results The Other Guys can DO stuff with.
>>
>> ...whoosh....
>
>   Perhaps ...
>
>   No, he never PAID for Python. Takes balls to bitch
>   about a free product  :-)
>
>   But I'm still confused what HE'S doing that seems
>   to get horribly depreciated SO quickly .....

Oh dear, your ignorance runs deeper than I imagined. Even with a hint
you don't see what is so amusing.

--
"First, find out who are the people you can not criticise. They are your
oppressors."
- George Orwell

"26C.Z969" <26C.Z969@noaada.net> writes:
> On 12/16/22 4:24 AM, The Natural Philosopher wrote:
>> On 16/12/2022 04:26, 26C.Z969 wrote:
>>> On 12/15/22 5:10 AM, The Natural Philosopher wrote:
>>>> On 15/12/2022 06:43, 26C.Z969 wrote:
>>>>>> The point is that I can update Bash and not worry about lots of
>>>>>> changed/depreciated stuff breaking existing scripts. It also avoids
>>>>>> trouble with scripts found on the web failing because they're a
>>>>>> few years old.
>>>>>
>>>>>    I still can't figure out what you're doing that
>>>>>    gets depreciated so fast ......
>>>>>
>>>> Especially as you never paid for it in the first place...
>>>
>>>    Well ... I'll give some credit to "hobbyists", but
>>>    yea, when you do it for a LIVING you can't always
>>>    be such a "purist". The Boss not only wants results
>>>    but results The Other Guys can DO stuff with.
>> ...whoosh....
>
> Perhaps ...
>
> No, he never PAID for Python. Takes balls to bitch
> about a free product :-)
>
> But I'm still confused what HE'S doing that seems
> to get horribly depreciated SO quickly .....

Python 2.x did not get deprecated quickly in absolute terms - 12 years
warning, in the end. However there’s an awful lot of 2.x code out there
and much of it has little or no maintenance effort available (either
because not enough people care, or because the people who do care have
other priorities). Forward-porting is not particularly difficult but it
does take time and effort.

Incidentally, ‘deprecated’ not ‘depreciated’.

--
http://www.greenend.org.uk/rjk/

On 12/17/22 2:08 AM, The Natural Philosopher wrote:
> On 17/12/2022 06:29, 26C.Z969 wrote:
>> On 12/16/22 4:24 AM, The Natural Philosopher wrote:
>>> On 16/12/2022 04:26, 26C.Z969 wrote:
>>>> On 12/15/22 5:10 AM, The Natural Philosopher wrote:
>>>>> On 15/12/2022 06:43, 26C.Z969 wrote:
>>>>>>> The point is that I can update Bash and not worry about lots of
>>>>>>> changed/depreciated stuff breaking existing scripts. It also avoids
>>>>>>> trouble with scripts found on the web failing because they're a
>>>>>>> few years old.
>>>>>>
>>>>>>    I still can't figure out what you're doing that
>>>>>>    gets depreciated so fast ......
>>>>>>
>>>>> Especially as you never paid for it in the first place...
>>>>
>>>>    Well ... I'll give some credit to "hobbyists", but
>>>>    yea, when you do it for a LIVING you can't always
>>>>    be such a "purist". The Boss not only wants results
>>>>    but results The Other Guys can DO stuff with.
>>>
>>> ...whoosh....
>>
>>    Perhaps ...
>>
>>    No, he never PAID for Python. Takes balls to bitch
>>    about a free product  :-)
>>
>>    But I'm still confused what HE'S doing that seems
>>    to get horribly depreciated SO quickly .....
>
> Oh dear, your ignorance runs deeper than I imagined. Even with a hint
> you don't see what is so amusing.

Do I need to study Zen koans for a few decades
first so I'll "get" whatever "it"/"the hint" is ? :-)

Certainly *I* never paid for his Bash-abusing scripts
(or is that "abusive Bash scripts" ?) but I'm also
wondering who *would*.

Software is a popular "hobby", but it's also a
paying occupation. The rules between the two are
not quite the same.

On 12/17/22 4:14 AM, Richard Kettlewell wrote:
> "26C.Z969" <26C.Z969@noaada.net> writes:
>> On 12/16/22 4:24 AM, The Natural Philosopher wrote:
>>> On 16/12/2022 04:26, 26C.Z969 wrote:
>>>> On 12/15/22 5:10 AM, The Natural Philosopher wrote:
>>>>> On 15/12/2022 06:43, 26C.Z969 wrote:
>>>>>>> The point is that I can update Bash and not worry about lots of
>>>>>>> changed/depreciated stuff breaking existing scripts. It also avoids
>>>>>>> trouble with scripts found on the web failing because they're a
>>>>>>> few years old.
>>>>>>
>>>>>>    I still can't figure out what you're doing that
>>>>>>    gets depreciated so fast ......
>>>>>>
>>>>> Especially as you never paid for it in the first place...
>>>>
>>>>    Well ... I'll give some credit to "hobbyists", but
>>>>    yea, when you do it for a LIVING you can't always
>>>>    be such a "purist". The Boss not only wants results
>>>>    but results The Other Guys can DO stuff with.
>>> ...whoosh....
>>
>> Perhaps ...
>>
>> No, he never PAID for Python. Takes balls to bitch
>> about a free product :-)
>>
>> But I'm still confused what HE'S doing that seems
>> to get horribly depreciated SO quickly .....
>
> Python 2.x did not get deprecated quickly in absolute terms - 12 years
> warning, in the end. However there’s an awful lot of 2.x code out there
> and much of it has little or no maintenance effort available (either
> because not enough people care, or because the people who do care have
> other priorities). Forward-porting is not particularly difficult but it
> does take time and effort.

So long as you can still get P2, a lot of that legacy
code is almost "immortal" now, they're not going to
be changing anything. You'll more likely be the victim
of underlying OS changes that make more and more library
routines obsolete.

Yes, it *does* require a little effort to do P2->P3, but
not an insane amount. Consider it an "opportunity" to
review/improve yer old code. IMHO, reviews should be done
every few years anyway - you WILL see ways to code it better.
My guess is P3->P4 won't be as bad.

Anyway, as I said, Python is not the ONLY way to take that
step above hideous abuse of Bash. It's an immensely popular
language and has a library or two or three for *everything*
which is a BIG plus. Perl is good too, even TCL can get
you there - and don't forget the good old c-shell. All
are more versatile and comprehensible than Bash taken to
its ultimate gobbledegoop extreme.

> Incidentally, ‘deprecated’ not ‘depreciated’.

DEP-re-cated ? Doesn't sound right. "Depreciated", as
in "losing value", better describes the situation :-)