4 years in consulting. I've spent the first WEEKS of a project twiddling my thumbs waiting for a laptop, just to spend more weeks waiting on access to source code, tooling, etc.
My friends on the strategy side could start and finish entire projects in that time.
Unless that was a different project altogether?
I've gotten so bored at work lately I've been coding for fun again
After recieving properly formatted email, script was executed to apply git merge between svn branches. In case of merge issues, the email was sent back with feedback. If everything was okay, a proper sign-off blessing by one of the technopriests as late check was applied and merge concluded.
I might be stupid, are you saying a build would take 9 to 24 months to finish?
Either its the wrong unit (minutes?) or the wrong definition of "build"?
It was really fun using filters in Pegasus Mail (no SOAP) to automate mailing lists, PGP key signing with e-mail validation etc.
Knowing I was a data engineer, one of the archaeologists asked me to take a look at the cataloging system he’d cobbled together on his own: a shared-drive Access database with a full-featured CRUD interface that the whole office had been using for years.
I was able to clean up one stray bug he had, and confirm his suspicion that one particular action was running slow because it had to touch multiple files by necessity (he’d rolled his own sharding) — but generally speaking, it was a work of art more effective than anything I could’ve ever come up with. Sometimes the “dirty hacks” are the best solutions.
But I have seen the maintenance burden first hand of solving weird Access lock file problems (if I never have to manually find and delete an .LDB file again, that would be great) and silent corruption issues and more. I've seen the workarounds of auto-backups of the shared folder and then auto-restores of those backups when silly things happen like the .MDB file is not the expected file size.
There's a special "joy" in needing to know the many under-the-hood versions of Access files and seeing apps that consume and/or produce more than one version at a time. That's just to maintain existing "apps", trying to migrate that data to modern databases for new apps is its own "joy" as well.
That's cute.
But it seems like an enormous security hole, even with a codeword "password". The author didn't mention it, but I hope they're using whatever version of their company's E2E email encryption is for these messages.
Ultimately the problem is that in a lot of big corps, IT is basically unaccountable for setting things up wrong. Their only KPI is tickets closed, not the quality or success rate of their fixes.
They default to tickets closed, uptime, SLA adherence as KPIs because you can't effectively measure "is it set up correctly?" and because the business absolutely must measure everything, they come up with bullshit KPIs so they can have a pretty dashboard and pretend like they're actually managing.
Glad I'm no longer in huge corps, but still an IT manager. Shadow IT is a direct symptom of IT not providing the right tools or having poor processes. But responsibility still lies higher up in the chain. If we weren't forced to quantify all activity, these issues wouldn't exist.
Hacky is as hacky does.
docker build -t your-docker-registry/xxx:latest .
docker push your-docker-registry/xxx:latest
kubectl rollout restart deployment xxx-deployment
Totally unrelated, if you need to suffix a k8s deployment with "-deployment" you are a hack, a fraud, an overengineer and someone who should NOT work as software or devops engineer.
You aren’t running tests, unless you put them in the dockerfile which is a bad idea…
You aren’t running security scans. how do you deploy manifest changes? Using Latest as a tag has so many issues.
This is a trivial and niave pipeline I would expect from a junior or intern.
Build pipelines are becoming more complicated because software is more complex. You can still promote ownership of the full pipeline while giving developers control.
Don’t shy away from it, understand it, embrace it. It’s just going to continue getting more complex
Sure, "docker push" is all fine and well until "after two weeks, [your] coworker still does not have access to the server endpoint that he and [you] would need". And then what? Do you quit your job for fear that someone calls you a hack?
go build binary
scp binary root@server:/deploymentlocation
you are a hack, a fraud, an overengineer and someone who should NOT work as software or devops engineer.
We had that at my first job where I deployed .war to a remote production tomcat directly from eclipse with one tool button click.
Zero lines of code even, depending on definition.
If you're doing anything else, are you really an engineer?
bwahaha.