What is it with our industry?

[Gibbon]

Hey, don't diss the macro. It's gonna be a work of art.

Hey, I have no doubt your 'macro' is a work of art, the Banksy of Excel!

[TheGreenBastard]

I concur with a lot of what the OP said, it's simply commoditization of the developer role, more and more responsibility is placed on their shoulders via nefarious schemes like "the Agile manifesto" (whether that's better or worse than waterfall I'm not sure). Developers are their own worst enemy, lots of hubris derived from a somewhat technical profession makes us believe we're "smart", developers themselves called for Agile, open plan offices etc. easily coerced into thinking this was their original thought by "management". I've frequently had developers praise Agile and open plan offices, true believers.

BTW OP, might want to research more if deriding a technology: https://www.prisma.io/dataguide/mana...tabase-caching

[mogga71]

I have taken some time to reflect recently on our industry. I usually use the time between contracts to consolidate knowledge, catch up on technical reading, write articles that are long overdue etc but this time I have had a much narrower focus. Pain. My pain. The pain caused to me by my industry, by its self imposed doctrines of the agile method, or at least their diluted and "perverted to fit JIRA" version of it, of MVP's and and POCs and scant regard for the baby or the bathwater.

I frame my pain not in the terms of the quality of the output, but rather the pain of getting there. The developer experience is, for me at least, an endless treadmill of hastitly written JIRA tickets with such malodorous vagueries as "etc". The levers of power that were originally deigned to be equally distrubuted among the self organising cross functional team are rested soley in the hands of the self agrandising "PO", Product Owner, who conveniently steps over the terms "Servant Leader" to just become "Leader". Woe betide anyone who asks what is the "definition of ready" for one of these JIRA tickets. No such impertinence is not tolerated. The powers that be have JIRA, the single source of truth, and you have your IDE and github. You type code, the code goes into the pipeline and good things come out. If not, then it is the developer's fault!

So imagine my stomach churning, gut wrenching surprise when I take the time to read the stack of tomes on the subject of agile development during my most recently arrived bench time, to discover most of it to be well thought out! That the problems with agile development are well known, and the kinks have been ironed out. The problem of the flat backlog should be no more, lest we rely soley on JIRA to structure our project into a blizzard of tickets that entropy into oblivion with every passing second.

And it is not just the crappy tickets that entropy, the malaise has spread to the industry as a whole, it seems. We seem to have gone fully over to the side of sloppy standards of requirements gathering, system design and documentation. The ability to learn empirically from our mistakes is taken away. The ability to refactor our way to safety stifled in the name of "velocity", or it's poorly understood version of it.

So far removed from the user base are we developers now that I cannot remember the last time I spoke with, let alone had a meeting with, end users. This vital connection has been severed, and in it's place is the conduit of the PO, whose lack of technical prowess constricts the flow of information to a trickle and yet the status quo must not be questioned.

This malaise seems to have met the perfect storm of new industry intake, who since they have over 3 years experience in such and such a group of in vogue technologies have the temerity to call themselves "senior" and the job roles that fill the job boards permit this by going along with this definition. This is juxtaposed with the speed and breadth of new technologies, such that anyone with less than say 10 years experience would be mostly working from stack overflow articles and best guesses unless they had the good fortune of spending their formative permie years in such an organisation as permitted them the time and resources (and perhaps training) to master these technologies and become expert to the benefit of the organisation.

I considered for a long time that I was just getting old and crotchety. A dinosaur that should have extincted myself into the mediocrity of middle management a long time ago. Alas I can't. I do what I do because I am driven by a thirst to learn and master, and a desire to do a good job, the kind that is met with happy(ish) end users. And my most recent contract I walked into a situation that really woke me up and drove home the point. Permit me if you will to sketch out the particulars for your amusement.

I was hired as a react/node/typescript coder. The interviewer pushed me on topics of Kubernetes, Docker, Networking, Advanced Typescript. I was informed I was "head and shoulders above the rest". High praise indeed. I got the gig. There was no Kubernetes or Docker or Networking or Advanced Typescript required for this role. I was tasked with "Load testing" an application that was due to go live in around 4 weeks. While the developers put the finishing touches to their product, and ironed out the remaining edge cases I set to work. But first I wanted to take a wander through the code base to see if I could glean any immediate hints of where to hit. Which routes were most likely to be hammered, prioritise those. Then suddenly I saw code for Prisma. I froze. Was this api actually connected to a Postges box? You bet your sweet bippy. Suddently I get that sinking feeling. How did they intend to deliver these millions of page hits?

So I contacted the Technical Lead with some questions. The fresh faced youth was unaware of the limitiations of horizontal scaling that beset all RDBMs'. From his perspective you wrote code, it hung together and all was well. I was fear mongering. OK. So I prepared some load testing and delivered some results. He pooped the bed. Then pestered me day and night on slack on how to deliver this. I liked the guy, and really wanted to help, and did. It went live with some of the fasted written caching software and message queueing goodness I could apply without requiring a rewrite. It scraped though it's load testing and went live. The security was lax, the lack of knowledge of OAuth2 grants and hastily smooshed in typescript libraries such as Passport to try and paper over the cracks did little to remedy this.

All of this was abstracted away in a haze of JIRA tickets. The tickets move to the right and all is well. The developers are fully utilised, all is well. The entire craft has been distilled into a noxious smelling quart of liquid which is then poured sloppily into a pint pot. The TL;DR; brigade march on with their thumbnail image of the world or 140 character precis.

Has anyone else experienced this slow burn descent into the depths of sloppyness and shallow knowledge? Why aren't the newspapers filled with news of failed big ticket projects?

Funny you mention the database issues and complete numpties who overlook them. In my experience, the majority of projects that do fail are down to issues with the underlying database. Companies hire 'full stack' devs who are supposed to know how to develop databases. Development on databases that scale is difficult and far harder than the sexy front end stuff. Only this week I had an interview with a team of full stack devs who were 2 years into a project and were experiencing horrible performance database problems .... and lots of tables without any indexes, no stored procedures .... no real sql devs. When they asked me about how I would fix issues I prattled on about SARGS, param sniffing, aligning data types, not using functions on predicates etc. It was only about an hour after the interview that it occurred to me that they probably didn't have a scoobie about what I was talking about but were too polite to ask. I won't be taking the job regardless. Just too much effort.

[vetran]

Funny you mention the database issues and complete numpties who overlook them. In my experience, the majority of projects that do fail are down to issues with the underlying database. Companies hire 'full stack' devs who are supposed to know how to develop databases. Development on databases that scale is difficult and far harder than the sexy front end stuff. Only this week I had an interview with a team of full stack devs who were 2 years into a project and were experiencing horrible performance database problems .... and lots of tables without any indexes, no stored procedures .... no real sql devs. When they asked me about how I would fix issues I prattled on about SARGS, param sniffing, aligning data types, not using functions on predicates etc. It was only about an hour after the interview that it occurred to me that they probably didn't have a scoobie about what I was talking about but were too polite to ask. I won't be taking the job regardless. Just too much effort.

I can see why they were confused

https://www.urbandictionary.com/define.php?term=sarg

Sadly few people appreciate large databases are fast because experts design them that way. Lets run our DW on 12GB with no indexes or any data remodelling.