this post was submitted on 30 Jan 2024
844 points (98.2% liked)
Technology
57313 readers
3748 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
If it ain't broke, don't fix it.
Until it becomes obsolete, unsupportable, the crux of your operation, and/or the basis for all of your decisions π¬
(Yes, I read the article, itβs just the signs, but yes, the above still applies!)
COBOL has entered the chat
e: good for legacy employment though. A relative of mine is a Z80 programmer by trade, and he can effectively walk into a job because the talent pool is so small now. Granted - the wages are never great but never poor, and the role is maintenance and troubleshooting rather than being on the leading edge of development - but it's a job for life.
Every time I hear about COBOL I feel like I should try to learn it as a backup plan...
I'm in two minds about that. One the one hand, yes, of course - as all the original COBOL folks die off, the skills will be even rarer and thus worth more.
On the other hand, if we keep propping up old shit, the businesses will keep relying on it and it'll be even more painful when they do eventually get forced to migrate off it.
On the other other hand, we know it works, and we don't want to migrate everything into a series of Electron apps just because that's popular at the moment.
Part of the problem is the cost of moving off it. Some companies simply can't pay what that would cost, and that's before you consider the risk.
Tough spot to be in.
You have to unlearn everything you know to learn it, go look its bad.
Let COBOL die, it's terrible.
If it works, why would we want to go through the trouble of switching to another language that will also eventually be regarded as needing to be retired? There's decades of debugging and improvement done on their system, start over with a new system and all that work needs to be done again but with a programming language that's probably much more complex and that leaves the door open to more mistakes...
I'm all for that I just never personally liked COBOL.
It wasn't for me, too wordy and felt more like something for accounting/corporate than a programmer. I was offered a good-paying job programming COBOL out of college but turned it down because I didn't want to spend my life with it. But that's just me.
π₯²
Not to mention when you want to change the entire system it becomes a huge operation and problem.
Massive risk to that change too.
So many people don't understand how risk informs everything a business does.
What cost is there to a given system being down for one hour? A day? Any regulations around it?
Often it's better to pay a known quantity up front than risk potential outages where you can't predict all the downstream affects.
I'd consider those various states of not working. So... Don't fix it if it's not broken!
Too critical to be upgraded is something I wish I'd never hear or see again in my professional career.
ha, whats the failover look like???
Manual and boot a machine with the same IP somewhere else. Very robust
Oh, everyone who ever travels by train in Europe will tell you that the German infrastructure is very much broken. You're lucky if your delay is less than a day travelling through Germany.
Well I live in germany and therefore use the train network on short and long distance frequently and while it is unreliable, "a day" of delay is something I have never experienced.
Most of the delayed trains are late by less than one hour (still atrocious, but not a day's worth by any means).
I actually experienced only once a situation where we were given the choice of a hotel or a continuation of our travels by taxi (which we chose) because the train we were in was late one hour or something and the other (last for the day) train could not wait.
Well, it's based on experiences travelling through Germany proper - for example Denmark to France or Italy, including transfers. Often the delay will just be a couple of hours, but then you miss your transfer and you're screwed.
Also if you're on your way to Switzerland the Swiss have no patience for disruptions in their services, so if a train is delayed coming from Germany they're likely to just not accept it into the country at all.
I have also heard from people who were told to spend the night in the train, which DB just parked in the outskirts of the city for the night. That way they could offer passengers a place to sleep in the cheapest possible ways. Pregnant women or families with young children were asked to check in to hotels.
Afaik Deutsche Bahn loses 5 mil a day: https://www.theguardian.com/business/2023/oct/14/its-the-same-daily-misery-germanys-terrible-trains-are-no-joke-for-a-nation-built-on-efficiency
Germany doesn't really seem like a very efficient country, they still use fax for things and every person has to manage like 10,000 different insurances for everything. Seems like an old (and inaccurate) ww2 trope.
It's mostly a misunderstanding of what is valued in German society. The common trope is that German society covets precision. This is not the case. German society covets unwavering precision in the adherence to norms. To the point where innovation is akin to revolution in the negative sense, and pigheadedness in procedure is considered a workplace virtue. In the mean time nothing gets done. Source: expat in Germany.
Is this the same as a migrant?
Yes, as long as they're also white and middle/upper class!
No, the way it's often used is closer to "posh guest worker".
Migrant implies the non permanent kind because a permanent migrant is referred to as an "immigrant".
What's the technical difference between a migrant and an expat?
Also bureaucracy is through the roof in everything, i have no idea who the fuck thinks of germany as efficient.
As an outside observation, Germans seem to make things better than they need to be in a detrimental way. For example, we redid one of our bathroom showers using the Schluter Kerdi waterpoofing system. They have very specific instructions on how to space the screws, how to seal the screws, how to seal the edges, how to mix the thinset, and probably some other things I can't remember off the top of my head. They put it through a battery of tests, including going under 100' of water. Who needs that? Don't worry about it.
This stuff replaces cement board, which isn't strictly waterproof, at least not on its own. It's also significantly more expensive.
I do think it's worthwhile for a home DIYer to get. The instructions are clear and it's less likely you'll screw something up that could result in disaster. That said, this thing is just waiting for a Japanese company to come along and make something 90% as good for 50% of the price. That's basically what happened in the German vs Japanese car market, and there's already some products on this market like that.
An old mechanic friend of mine used to say "German cars are over-engineered and under-designed", lol.
Having worked on every brand of car out there, his description, and your explanation make a lot of sense together.
I've never seen such a clear and concise comparison of German/Japanese manufacturing, you really nailed it.
Both approaches have their place, the key is to know when to apply them.
I have no doubt their bureaucrats perform world-class efficiency in their handing out, filling in, faxing and archiving a sophisticated system of paper forms.
I guess it's the trap of getting complacent and stopping modernizing as soon as you've convinced yourself you have the best system in the world.
It's more that the bureaucracy is so complex and fragmented that it's incredibly hard to digitalize. Lots of small fiefdoms that are entitled to make IT purchasing decisions themselves means paper is the only universal interchange format. In addition there is an unwillingness to change how things have always been done, or to simplify procedures. So there you have it: The German bureaucracy is too fat to move.
I work for german government agencies from time to time and they are working on it... It's just really slow because there is so much of it, and due to organizational overhead. Also, there is not a single push for the entirety of Germany, but some things everyone does for themselves.
He's actually German.
German re-unification cost trillions. It's entirely unsurprising.
That's another part of the infrastructure, though: We just don't have enough rail as well as backup rolling stock.
And as the federation finally decided to spend some money it's going to get worse in the next decade or so due to outages due to new constructions being linked up to the old stuff.
As to the age of the infrastructure -- I mean it's the railway. If a rarely-used branch line still uses mechanical interlocks and there's no need to upgrade the capacity then the line is going to continue using infrastructure build in the times of the Kaiser. It's not like those systems are unsafe, it just might be the case that unlike in the days of ole those posts with a gazillion levers aren't manned all the time so you'll see an operator drive to it with a car while the train is on its way. Which really isn't that much of a deal when the branch line goes to a, what, quarry maybe sending out a train every two months or so. Certainly better than to demolish the line and use trucks instead.
Is it broke if no one is able to fix it?
The reason for it to run on such an ancient device is because nobody wants to touch the scripts running on these devices.
A lot of these systems are also always on.
Used to work at an airport that had a similar issue, turning some of these systems off simply isn't possible. So you end up having to run the replacement system simultaneously with the old system for a few days. Can't simply take it off line for a day.
Running two systems simultaneously for a couple of days, that's a huge problem, not solvable
It's an expensive problem, especially if it's a system that's being used all across the airport by regular staff.
You need to train thousands of employees to use the new software, you need to have one person using the old software as a backup, while the other uses the new software, often while surrounded by hundreds of often angry customers.
And if something goes wrong, which it invaribly does (even if it's user error or someone snagging a cable), shit can get very expensive. Small delays, add up to larger delays, and cascade through the entire system. Delayed flights, tens of thousands of euros in costs, hotels for thousands of passengers, missed flights, missed meetings, damages, lawsuits, penalties for missed landing/take-off slots, missed time windows for certain cities which don't allow flights after a certain time, etc. And often you discover legacy stuff while you're upgrading that needs fixing, stuff that no one knows how to replace anymore or is physically hard to access.
Sometimes it is genuinely better to leave it. COBOL is 60 years old. There's still plenty of stuff running on it, exactly because it's often too expensive and too risky to replace.
These probably operate completely shut from other networks/internet, so I definitely agree. But I guess a lot of folks here are Linux maniacs and can't stand something running ancient and obsolete OS while the all-mighty Unix-based operating system could solve all of the problems, not mentioning that it would create more in the process.