Older contractors

[d000hg]

I think in 5-10 years only the old programmers will be working looking after legacy systems. The new generation will ask a daft AI to write a system for them and it will probably do a decent job.

I have "only" been in development/coding for 20 years but even before I graduated, people were prophesying "the end of coding". As it turns out, not only has coding not been replaced by visual tools and code generators, it's hardly changed. The tools are far, FAR better but TBH even the languages haven't changed much. Web is still written in JS, backend still often written in PHP/Java... .Net is great but evolutionary from Java/C++ and Google et al are still trying to find a replacement to C++. New tech and languages come and (mostly) go but the day to day coding work is really not massively different.

Maybe AI is the disruptor everyone says - I am leaning towards that view - but complex engineering in niche areas with bespoke terminology are not an easy application for it. "Write me code to X" is all very well, but figuring out precisely what "X" is is something humans haven't really figured out. A new developer joins a team, even if they are great they have no understand what the code does. They have to scrutinise code, documents, etc and often talk to the rest of the team... I'm not sure if that part of AI/ML is as far along as the 'flashy' bits.

[TheDude]

I think in 5-10 years only the old programmers will be working looking after legacy systems. The new generation will ask a daft AI to write a system for them and it will probably do a decent job.

I remember a lecturer at UNI saying that being a clerk in a bank was a middle class profession 100 years ago with status until tech replaced the need for any kind of skill. I can see general programming going the same way.

It will be those who refuse to or are unable to adapt that will fall by the wayside.

Those crusty fossils who ridicule new technology and hide their incompetence behind bluster and war stories.

[woohoo]

I have "only" been in development/coding for 20 years but even before I graduated, people were prophesying "the end of coding". As it turns out, not only has coding not been replaced by visual tools and code generators, it's hardly changed. The tools are far, FAR better but TBH even the languages haven't changed much. Web is still written in JS, backend still often written in PHP/Java... .Net is great but evolutionary from Java/C++ and Google et al are still trying to find a replacement to C++. New tech and languages come and (mostly) go but the day to day coding work is really not massively different.

Maybe AI is the disruptor everyone says - I am leaning towards that view - but complex engineering in niche areas with bespoke terminology are not an easy application for it. "Write me code to X" is all very well, but figuring out precisely what "X" is is something humans haven't really figured out. A new developer joins a team, even if they are great they have no understand what the code does. They have to scrutinise code, documents, etc and often talk to the rest of the team... I'm not sure if that part of AI/ML is as far along as the 'flashy' bits.

I agree, I’ve been a paid as a programmer for 30 years now. But i can feel a sea change.

Perhaps 5 years is too soon but 10 years mmm think thats about right.

[DoctorStrangelove]

I have "only" been in development/coding for 20 years but even before I graduated, people were prophesying "the end of coding".

Ain't they just:

https://en.wikipedia.org/wiki/The_Last_One_(software)

.

It was The Coming Thing until it wasn't.

[The_Equalizer]

Also using thumbs up is old. You are supposed to send a facial expression emoji.

Breaking the thumbs up habit is hard, although Boris and Paul McCartney spring to mind every time I do it.

[vetran]

I was partly joking but I'm not sure, I can see in 5-10 years there being a big change in general programming. But I'm just guessing. I can still remember a bit of Cobol, I don't think a decent AI worth its salt would lower itself to that so I should be ok Jack.

I think someone needs to teach the AI which stones to turn over, that normally is the expertise. i.e. which incredibly stupid thing did the coder do, and you know how the Universe likes inventing a new idiot

[ladymuck]

I am in the downward trajectory to my fifties (my Mum yesterday was being shocked at how old her children are!) and have been bimbling along in contracting for 17 or so years.

Never had a plan, couldn't tell you what I do (it's not secret, I just have never fathomed what my elevator pitch is), yet the work still turns up. I do dye my hair but don't wear skinny jeans. I do say 'cool' too often.

[Paralytic]

I was partly joking but I'm not sure, I can see in 5-10 years there being a big change in general programming. But I'm just guessing. I can still remember a bit of Cobol, I don't think a decent AI worth its salt would lower itself to that so I should be ok Jack.

IDENTIFICATION DIVISION.
PROGRAM-ID. COUNT-GOOD-POSTS.
ENVIRONMENT DIVISION.
INPUT-OUTPUT SECTION.
FILE-CONTROL.
SELECT POSTS-FILE ASSIGN TO "posts.dat".
DATA DIVISION.
FILE SECTION.
FD POSTS-FILE.
01 POSTS-RECORD.
    02 POST-ID PIC 9(4).
    02 POST-USER PIC X(20).
    02 POST-QUALITY PIC X(3).
    02 POST-DATE PIC X(10).
WORKING-STORAGE SECTION.
01 WS-COUNTER PIC 9(4) VALUE ZERO.
01 WS-USER PIC X(20).
PROCEDURE DIVISION.
BEGIN.
    DISPLAY "Enter the user name:".
    ACCEPT WS-USER.

    OPEN INPUT POSTS-FILE.
    READ POSTS-FILE
        AT END SET WS-COUNTER TO ZERO
        NOT AT END
            IF POST-USER = WS-USER AND POST-QUALITY = "GOOD"
                ADD 1 TO WS-COUNTER
            END-IF
           READ POSTS-FILE
    END-READ.
    CLOSE POSTS-FILE.

    DISPLAY WS-USER "has made" WS-COUNTER "good posts".
STOP RUN.

[Snooky]

IDENTIFICATION DIVISION.
PROGRAM-ID. COUNT-GOOD-POSTS.
ENVIRONMENT DIVISION.
INPUT-OUTPUT SECTION.
FILE-CONTROL.
SELECT POSTS-FILE ASSIGN TO "posts.dat".
DATA DIVISION.
FILE SECTION.
FD POSTS-FILE.
01 POSTS-RECORD.
02 POST-ID PIC 9(4).
02 POST-USER PIC X(20).
02 POST-QUALITY PIC X(3).
02 POST-DATE PIC X(10).
WORKING-STORAGE SECTION.
01 WS-COUNTER PIC 9(4) VALUE ZERO.
01 WS-USER PIC X(20).
PROCEDURE DIVISION.
BEGIN.
DISPLAY "Enter the user name:".
ACCEPT WS-USER.

OPEN INPUT POSTS-FILE.
READ POSTS-FILE
AT END SET WS-COUNTER TO ZERO
NOT AT END
IF POST-USER = WS-USER AND POST-QUALITY = "GOOD"
ADD 1 TO WS-COUNTER
END-IF
READ POSTS-FILE
END-READ.
CLOSE POSTS-FILE.

DISPLAY WS-USER "has made" WS-COUNTER "good posts".
STOP RUN.

It's several decades since I last did COBOL, but back then different column regions denoted different functions of the code, e.g. Area A & B etc. Is this no longer the case?

[Paralytic]

It's several decades since I last did COBOL, but back then different column regions denoted different functions of the code, e.g. Area A & B etc. Is this no longer the case?

The COBOL language has not changed significantly since its inception in the late 1950s. The basic structure and format of COBOL code, including the use of columns for different purposes, is still in use today.

In COBOL, columns 1 through 6 are reserved for sequence numbers, columns 7 through 11 are typically used for indicators or continuation characters, and columns 12 through 72 are used for actual code. The columns are used to define the structure and organization of the code, with specific functions assigned to different regions.

The use of column regions is still a standard practice in COBOL programming, and it is used to improve the readability and maintainability of the code. While there may be variations in the specific uses of each column region, the basic structure of COBOL code remains largely unchanged.

(in case its not clear, these are ChatGPT's responses not mine - i've repasted the code above to show the formatting it gave)