On 2017-09-23 22:40, Victor Odhner wrote:

I just found time to catch up with PLUG, and now as an obsolete programmer (in recovery for 4.5 years) I have to ramble a bit.

 

But in a few words, my point is: human thought builds on the past. Of course, learning details of old skills may be interesting, but ceases to be profitable for most. We need to create new things, learn new ways to perceive and tackle bigger challenges. There are opportunities out there that we can't even imagine until we go for it.

 

In Heinlen's 1953 story, Starman Jones serves as a "computerman," flipping switches to enter astrogation data in raw binary. That's how they did things in 1953's future. Some of the early Star Trek tools and procedures became comical as reality overtook fiction. (I remember a programmer deftly entering binary addresses or commands, one octade with each swipe of his hand, to boot up a big I/O unit, back about 1969.)

 

First there was raw code, octal or hex. Someone invented the assembler, that made it more readable and supplied some macros. Then came the FORmula TRANslator (aka compiler), a giant step. My programming started with ALGOL, and I went through obscure languages including some Prolog, and ended with C, shell and Perl. I became a dinosaur as Java took hold — I had burnt out learning all that an asterisk means in umpteen languages. I made web pages in 1994, and now I can't read the source code of the enormous, multilayered machine-generated pages.

 

Where I now volunteer doing facilities chores, I just fell off my IT-recovery wagon to learn Access and SharePoint (yeah, they're committed to MS) to build a simple work-order entry and reporting tool. In my first hour of study I was informed by a Microsoft message box that we should move on to PowerApps instead, because it's mobile-savvy and has more layers of ready-made tooling. I wouldn't want to leave StreetLightUSA.com looking for the last decade's obsolete MS skills . . . .

 

---

I do not have the depth of your experience.  I was first exposed to programming in 1983 as an accounting Major at the UofA.  At that time business majors were required to take intro to computers which was programming in Fortran.  Accounting majors were required to also take a class in Cobol.  I liked Cobol and changed my major to MIS. 

 

Cobol in academia was structured and organized.  An internship with the city of Tucson exposed me to a different Cobol - a Cobol that was difficult to debug because it was spaghetti code with lots of go-to's and lacked structure. 

 

Those were the days of punch cards and single line editors.

 

You write of MS-Access... the only option for desktop database application development unless you add Java and dot net.  This is an undeserved niche.  I took a look at MS-Access recently and found you are required a license for each P.C. I suspect you can run it on a server and expose it to each client w/o needing more than one license.  I did not get that far into my research because I have been spoiled by FoxPro and open source.

 

I was a dBase developer for almost 15 years, programming with multiple dialects of dBase, until MS bought and killed off FoxPro. 

 

I believe (an assumption of mine) that there is a real need for something simple like Visual FoXPro to build custom applications for small businesses.  Then again I could be missing the boat because today everything needs to be able to interconnect with everything else.

 

Things have gotten much more complicated.  I have had to build new skills several times and am doing it again.  There is a developer in Scotland that I have followed and he says one should take every fourth year off for learning and skill building.  I think we need to structure our learning into weekly chunks of 5 - to hours.

 

I sure wish I could find something like VFoxPro...... I miss those days.   

  

---

 

 

I see a continuous growth of new layers, so that we no longer think about what lies beneath. Hardware logic was built with vacuum tubes, and now we have CPU chips with a lot of the same logic: they don't have to create much logic now, though a few brilliant people may be inventing new stuff; mostly they just use existing designs. Engineers became programmers, using a programming language to evoke known bits of logic into the design of chips.

 

Starting in 1890 my father's cousin Willgodt built Odhner Arithmometers (calculators) by the thousands, and millions of clones were produced worldwide until electronic calculators took their place. For my part, in about 1970, I did semi-automation to install 18,000 wires onto each 6x2 foot CPU back-panel; now the phone in my pocket puts that high-end system to shame. Everything shrinks, and gets more complex.

 

My PC has hardware logic, the BIOS, the operating kernel, the services of the OS, and applications such as web browsers. I can write straight HTML, but then they added javacode, but wait, there's more: layer upon layer of canned tools to save the webmaster from the need to create new code. Now lots of people can make web sites, complex ones, in a hurry. And it looks like there's a big demand, but not for hacking a lot of HTML.

 

Arthur C. Clarke publicized the idea of the geostationary broadcast and telecommunications satellite, and addressed the problems of power (a solar-driven steam engine) and maintenance (someone had to be there, to change failed vacuum tubes). He projected that the idea could be realized within 50 years. He didn't know that a modern photocell would be patented the next year, and transistors would emerge a year later. The first satellite-relayed phone call (JFK to the president of Nigeria) came in 1963. He was one of the brightest of us, a brilliant futurist, but he under-estimated how fast things are moving.

 

Dr. Edsger Dijkstra, a prominent computer scientist, visited us at the big Honeywell Informations Systems factory in Phoenix (now a Best Buy), for Engineers' Day, and gave us a talk about proof of correctness in programming. During the break, over a soda, I commented to him that my program code was simple minded, but most of my work was determining what results we wanted. He nodded, and after the break made a disclaimer with words to this effect:

The specification is a firewall between the pleasantness issue and the correctness issue. Correctness applies to the program code, and its correctness can be proven. But nobody can prove that what you asked for will be what you really wanted.

 

It's fun coding to the bare metal, but yes, it's mostly been done.

 

A programmer-analyst is a translator, converting requests and observations into an approximation of the desired automation. But those skills apply to lots of things, not just automation. The involve listening, observing, thinking logically, giving a hoot, and being willing to step outside the box.

 

Programming as we know it will be obsolete, but the sky's the limit.

__________________

 

On Sep 14, 2017, at 12:59:50, David Schwartz <newsletters@thetoolwiz.com> wrote:

For now, I think that IFTTT and Zapier represent the "leading edge" of where programming is heading for lots of otherwise routine needs.

 

With the growth of SaaS, it's becoming more of a problem of simply "wiring things together" than programming.

 

I'd say that well more than half of my coding is simply what I call "plumbing", and the only reason it takes so frigging long is that people are still overly concerned about "efficiency" in areas where it's virtually irrelevant.

 

Back in the early 80's I had lots of quite vehement arguments with OS guys about how the only real way to solve some of the problems we were facing was to send raw ASCII data with self-identifying flags over the wire, and they'd scream back that this was simply too inefficient, and we had to come up with a bazillion "cannonical data tables" to define every possible common string we could think of so we only needed to send a one or two-byte value instead of a string. 

 

They were of the school, "Hey, do you guys remember joke #729? What a whopper that one was, eh?!"

 

HTML came on the scene and us old farts almost choked over the fact that it was stateless and you had to send the entire freaking page back and forth, often multiple times, just to do the simplest things.

 

Here we are 20 years later and javascript has finally overcome the stateless nature of web pages so now we can send several KB of data back and forth between the client and the server without updating the screen, but those hidden transactions end up moving far more data in many cases than what it takes to refresh the entire screen. And caching has gotten a lot more refined, which reduces the need for data to move around.

 

Anyway, we're going to see more ground taking along the lines of making common problems solvable without programming. This is what IFTTT and Zapier are doing, among others.

 

The tradeoffs will be time-to-market vs. where a new requirement falls along a dimension of "common/generic vs. fully custom" in terms of UI, UX, and logic.

 

At some point, we'll see voice-controlled solution drivers that assemble IFTTT and Zapier in the background!

 

This is NOT "programming". It may end up replacing a lot of Really Boring Jobs, but who cares? Is this the kind of stuff YOU want to do as a programmer? I sure don't!

-David Schwartz

 

On Sep 14, 2017, at 9:23 AM, techlists@phpcoderusa.com wrote:

I just read this article  :  https://www.linkedin.com/pulse/learning-how-code-still-worth-rajat-bhageria which predicts that computers will be self coding and coding skills will be obsolete.

I've read other such articles in the recent past.  I'm also reading about robots replacing jobs.... And as a precursor sending jobs off-shore will become the norm.  I get contacted all the time by Indian programmers willing to work for 10% of what an American contractor is willing to work for.   Even though I never have, I know there is a lot of challenges of working with offshore Indian programmers... And I know there is a lot of challenges working with American employees...

---------------------------------------------------
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

---------------------------------------------------
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

---------------------------------------------------
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss