@thomasfuchs
You wax poetic about how much better things were in your day...
My brother in Christ, you're the CTO.
@thomasfuchs
You wax poetic about how much better things were in your day...
My brother in Christ, you're the CTO.
@thomasfuchs of course, when they actually do need to change something they find that, even if they have the source code, everyone who understands it is either retired, dead, or gone into consulting at 10k per day…
Underline that last line. Some just like to follow those shiny hot stuff blindly.
Exactly this logic is why it has taken me 2+ years to decide that, yes, it looks like #Rust is actually worth the time and effort to learn.
At this point, I'm pretty sure it is one of those languages that'll still be around 50 years from now.
PS: Still waiting patiently-ish for JavaScript to please just die already.
@thomasfuchs If they'd spent the last 10 years mastering COBOL instead of stumbling along with React, they'd be set for life.
@thomasfuchs as someone who graduated in the late 00s and straight into mainframe work. It's interesting to explain, yeah it's super stable and has crazy transaction volumes it can sustain (the history of IMS is a bit bonkers). Buuuuuut it's a weird working market of older developers refusing to knowledge share and younger devs being locked out of the space by mockery and lack of teamwork and on the job training.
It's going to get real weird when those older devs start expiring.
@thomasfuchs listening to a room at IDUG clap and cheer at a story of a new dev not getting any training, following a manual to the letter causing a bigger problem by bad documentation...then the story turned to a name and shame which made those jackals laugh showed me that the industry there has systemic problems about knowledge hording and it's going to sting hard very soon.
@thomasfuchs While there is value in stability, that’s not why banks still run on the mainframe. It’s just cost/benefit analysis: it’s cheaper to live with all the technical debt than it is to figure it all out and rebuild it. COBOL didn’t win because of any intrinsic value, it’s just where investment stopped.
@thomasfuchs Cobol is a brilliant language, and it is not easy to hack. It does what it says. I said in an interview recently that it was one of my favourite languages.
Because it is simple. It deals with mapped memory.
If it still works, it is fine.
@thomasfuchs yeah, that strategy doesn't always work that well. https://www.bbc.co.uk/news/business-64036529.amp
I've also worked on legacy systems that no one understands any more and there is no spec other than the live system. I think relying too much on something that people don't want to work in any more is a false economy. The TSB system before the migration was failing as well and they couldn't get anyone to fix it
@thomasfuchs
brace for the obnoxious RIIR comments coming...
@thomasfuchs Running 60 year old code benefits the businesses because they don't have to pay developer time to update their software to conform with modern development standards. When the code breaks, the bank will pay out to ensure the code is repaired in place instead of gradually updating components to ensure the code base does not go stale and conforms with modern development standards that junior developers can work with. Just because it's old doesn't mean it glitters.
@thomasfuchs Also, that behavior was pored over, specified, written to standard, tested by standard, and the misunderstandings were *themselves* re-specified. People say "waterfall" like it's a bad thing.
@thomasfuchs Having worked on many of these systems when I was consultant it's often 'the team who wrote it left and we have no clue how it works'. So no team / manger wants to touch old systems & the risk of building replacements without understanding the original is way too high for the super-risk averse banks. ind you it employs a bunch of VB6, COBOL and LISP devs who otherwise would be skilled out of the market so not all bad ;)
@thomasfuchs truth.
A neighbor of mine wrote & maintained a payroll system written in #fortran for 110,000+ employees until she retired after 50 years there.
Her last programming task was writing, migrating and documenting the replacement system she wrote when she was hired in C++
@thomasfuchs @donmelton I had a terse email exchange with the people behind a certain popular crm/blogging system a number of years ago. I wrote to complain that they didn't prominently link the release notes for each new release and it was a number of steps to find them. That made it a hassle to go look and see if the release covered issues I would have and thus made it worth installing and testing. They responded that "most of our users just want to run the most current thing.”🙄
@thomasfuchs When I worked in retail in my youth and every year we had the bright shiny new stuff, it taught me to greet every new feature with, "Yes, but what does it do?" It's amazing how often the answer is nothing.
Graphics and sound aren't really important with bank records, AFAICT.
@thomasfuchs if you are going to post boomer nonsense you can give us all a perfectly good reason why you aren't building new software in FORTRAN.
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.