Software upgrades in general

Joined
Jul 10, 2022
Messages
3,670
This morning, I got a popup as I was pulling some video from the dash cam micro SD, new version of the VLC media player available.

To be fair, I'm on Win7 Pro on this older laptop.

I download and install, it no longer works, at all. Crashes every time it's launched.

What was very cool is they have their archive available and so I thought back and know that I got this dash came around March 2023. I uninstalled, and installed the version from March and all is good.

Yes, I should have remembered that this machine is running Win7 (I declined a free upgrade on purpose--I have BMW software installed that does not run on Win10), so I should have not installed.

But, software in general that I can see from work, generally is full of bugs, yet there is always this pressure to upgrade to the latest/greatest. And when we do, problems.

I also think developers are under the gun to churn out new and higher numbers and revisions etc., and like everything, the quality is questionable.

Going all the way back to the above, for some reason, I could not play the videos from the dash cam the first time I tried. I searched online and VLC seemed to be a good player to choose, and it worked. And I've used it since....

Any good/bad software upgrade stories? I did brick a brand new Zebra ET56 tablet once by shutting it off when it was doing an update, had to send it back as it had both warranty and contract coverage...
 
You’re trying to run wildly mismatched revisions of OS and software. VLC works fine on newer Windows versions. No conclusions can be drawn from this specific experience other than the latest VLC is incompatible with your OS. That being said…

The only way fixes can happen is through updates. The only way new bugs can occur is through updates. Humans naturally tend to remember bad experiences and forget good experiences, so it appears to them like updates correlate with issues more than fixes. This is not necessarily the case, although if your premise is true it may be due to the following:

Software practices have changed from waterfall to agile over the past 40 years, with higher emphasis placed on fast, lightweight iteration and less emphasis on “do it all from start to finish.” When implemented incorrectly, I could see agile contributing to lower quality under the wrong circumstances. There are many things to dislike about waterfall, but at least the end result of lengthy stabilization is a decent (usually late and over budget) product.

What tends to make poor quality happen in agile is overcommitting scope, failing to be always shippable, and releasing too soon. These are easy traps to fall into. One advantage is that you can release more often, so at least you can fix your mistakes quickly… in theory.

Overall, agile is better than waterfall, but harder to do well. It requires discipline, acceptance of reality, and foresight. If pm/tech teams lack these characteristics, they tend to overcommit and deliver immature software.
 
Any good/bad software upgrade stories?

I've got an antiquated desktop running windows 10. Yesterday, it decided to do an update, so I clicked okay and I set the system to download and shut down by itself. I then shut the monitor off. It was taking quite some time, so I went back and I turned the monitor back on to see the status. It said that the upgrade had failed and it was returning all the settings to pre-upgrade. Whatever, it works okay now.
 
VLC is run independently from any operating system. I have it running on UBUNTU right now. Apparently those that do support VLC have been VERY busy lately since I have gotten many MULTIPLE upgrades to the program over the last 6 months.
The only issue I have is that there is not a detailed description of which features are being upgraded on Linux.
 
I've got an antiquated desktop running windows 10. Yesterday, it decided to do an update, so I clicked okay and I set the system to download and shut down by itself. I then shut the monitor off. It was taking quite some time, so I went back and I turned the monitor back on to see the status. It said that the upgrade had failed and it was returning all the settings to pre-upgrade. Whatever, it works okay now.
There is some credence to if it ain't broke don't fix it. What I do see in the corporate world is a fear of anything remotely related to security and vulnerability--upgrade! Just pointing out things that don't make sense. I have the ability to install software on my own laptop at work (local admin), yet the co. stated that's been taken away. Whatever. I have the ability to install on others' machines, and even I question should I? I don't work on desktops. imho all of this happens when a small co. becomes a medium co. So many loopholes exist and even when people know about them, they don't know where to begin.

Lately our ability to remote control to RF devices stopped working. A teammate discovered, hey clear your cache for 24 hours it works again! Sure, it does work. But all kinds of other things are now screwed up as a result (seemingly related to 2 factor authentication and deleting cache). So for one system (a MDM), others are now messed up....
 
There is some credence to if it ain't broke don't fix it. What I do see in the corporate world is a fear of anything remotely related to security and vulnerability--upgrade! Just pointing out things that don't make sense. I have the ability to install software on my own laptop at work (local admin), yet the co. stated that's been taken away. Whatever. I have the ability to install on others' machines, and even I question should I? I don't work on desktops. imho all of this happens when a small co. becomes a medium co. So many loopholes exist and even when people know about them, they don't know where to begin.

Lately our ability to remote control to RF devices stopped working. A teammate discovered, hey clear your cache for 24 hours it works again! Sure, it does work. But all kinds of other things are now screwed up as a result (seemingly related to 2 factor authentication and deleting cache). So for one system (a MDM), others are now messed up....
Security vulnerabilities are constantly being discovered and addressed, so remaining up to date is critical if you want to be most secure.

IT groups tend to get handed poor incentives, though. They are tasked to operate with the philosophy of “prevent security breaches at any non-dollar cost” and also “save expenditure.”

This creates huge issues for everybody else in orgs as IT locks down permissions, installs incredibly intrusive, resource-hogging, process interrupting security software on increasingly old hardware, migrates pools of users to overstressed shared cloud infrastructure, etc.
 
I work on embedded software and do circuit board layout in my work. I absolutely resist doing software upgrades, especially with our CAD tool for circuit board design. I'll upgrade from one version to another, and I'll find that they have changed menus around, removed or renamed features, and there are usually new "quirks" that creep in. One quirk was the tool that is supposed to find problems in your layout, like a short circuit, was broken. It appeared that the board was fine and passed all the error checks, but it turned out there was a short circuit. So, we had 500 boards arrive that didn't work until I found the short circuit. Fortunately we were able to carefully cut a break in the trace to fix all 500 of them, but when I got in touch with the support folks to report the problem, they said "Oh yes, we just fixed that problem, upgrade to the newest version.". They didn't think it was very funny when I asked them what will be broken in this newest version considering we had just upgraded to the version that had this problem.
 
I work on embedded software and do circuit board layout in my work. I absolutely resist doing software upgrades, especially with our CAD tool for circuit board design. I'll upgrade from one version to another, and I'll find that they have changed menus around, removed or renamed features, and there are usually new "quirks" that creep in. One quirk was the tool that is supposed to find problems in your layout, like a short circuit, was broken. It appeared that the board was fine and passed all the error checks, but it turned out there was a short circuit. So, we had 500 boards arrive that didn't work until I found the short circuit. Fortunately we were able to carefully cut a break in the trace to fix all 500 of them, but when I got in touch with the support folks to report the problem, they said "Oh yes, we just fixed that problem, upgrade to the newest version.". They didn't think it was very funny when I asked them what will be broken in this newest version considering we had just upgraded to the version that had this problem.
What were you expecting from them when you reported the bug? They gave you great news: a fix was already implemented. Because humans haven’t yet invented time travel, the only way to get a fix to you is to give you another update.

It’s pretty rude to respond the way you did to the best news they could give you given the circumstances.
 
What were you expecting from them when you reported the bug? They gave you great news: a fix was already implemented. Because humans haven’t yet invented time travel, the only way to get a fix to you is to give you another update.

It’s pretty rude to respond the way you did to the best news they could give you given the circumstances.
Maybe so, but we've found the quality control on this software is pretty appalling for something that costs us about $4500/year, and what they call "upgrading" is more like "roll the dice and hope things aren't broken" or "re-learn how to use the features we've been using because they've changed things enough but the documentation hasn't caught up yet". Browsing their support forums, there are lots of folks like me who figure they found a version that works, so don't gamble on a new version.
 
Maybe so, but we've found the quality control on this software is pretty appalling for something that costs us about $4500/year, and what they call "upgrading" is more like "roll the dice and hope things aren't broken" or "re-learn how to use the features we've been using because they've changed things enough but the documentation hasn't caught up yet". Browsing their support forums, there are lots of folks like me who figure they found a version that works, so don't gamble on a new version.
In my experience this is often as much an attitude problem as it is a software problem. A bad attitude is going to actively seek out things it doesn’t like every bit as much as legit problems are going to cause legit pain.

If you truly feel your frame of mind isn’t part of the issue (which without introspection is pretty easy), and they are really truly regressing their product, then I suggest dropping the software. If you’re unwilling to update, then the relationship is dead. Existing somewhere in between never goes well.
 
Security vulnerabilities are constantly being discovered and addressed, so remaining up to date is critical if you want to be most secure.

IT groups tend to get handed poor incentives, though. They are tasked to operate with the philosophy of “prevent security breaches at any non-dollar cost” and also “save expenditure.”

This creates huge issues for everybody else in orgs as IT locks down permissions, installs incredibly intrusive, resource-hogging, process interrupting security software on increasingly old hardware, migrates pools of users to overstressed shared cloud infrastructure, etc.
I think my employer got from 700 mil to 5 bil in revenue in a decade by being able to respond to customers’ needs faster than the competition. Now, there is an ongoing joke there are 9 VPs to 1 worker. A hiring freeze yet oh there’s another svp that joined.

Anyway, to illustrate. Some vendors’ software got flagged by crowd strike. Can’t use it. I’m not in infosec so I can’t speak to the path forward. The co spent 7 figures on that engagement and now it’s rejected by a “AVP.”

The organization using the software brings in revenue. It massively saves on labor costs. IT doesn’t make any money at all, it spends. So maybe the right thing is being done. But we’re losing money.
 
Maybe so, but we've found the quality control on this software is pretty appalling for something that costs us about $4500/year, and what they call "upgrading" is more like "roll the dice and hope things aren't broken" or "re-learn how to use the features we've been using because they've changed things enough but the documentation hasn't caught up yet". Browsing their support forums, there are lots of folks like me who figure they found a version that works, so don't gamble on a new version.
My former career was telephony. I’d like to think we were the best of the best, bordering on Bell Labs. 😊

But the service “contract” on telephony gear exceeded an FTE’s comp like mine. All cos looked to cut it and rely upon experts like I was. Almost none did. Flash forward to maybe 8 years ago. They simply got rid of the equipment and the crew who handled it and paid cloud providers instead ($8.50/mo per user which is cheap).
 
I work on high security Linux systems. Most of the time the problems I have in upgrading haven’t been written about nor experienced by tech support. Very frustrating and time consuming.

We’re very close to the point where programmers will be required to have a verified identity and have to digitally sign their software, or it won’t be allowed installation.
 
i hate the added bloat and UI/UX fisher-pricification of every new iphone update
 
I work on high security Linux systems. Most of the time the problems I have in upgrading haven’t been written about nor experienced by tech support. Very frustrating and time consuming.

We’re very close to the point where programmers will be required to have a verified identity and have to digitally sign their software, or it won’t be allowed installation.
Did you here about the Linux recall? Someone uploaded hate speech into an interpreter module. Have to produce a do-over.
This is a recent distro issue.
 
Did you here about the Linux recall? Someone uploaded hate speech into an interpreter module. Have to produce a do-over.
This is a recent distro issue.
I haven’t read about that.

I’m trying to get through https://www.ca5.uscourts.gov/opinions/pub/23/23-30445-CV0.pdf

State of Missouri; State of Louisiana; Aaron Kheriaty; Martin Kulldorff; Jim Hoft; Jayanta Bhattacharya; Jill Hines,
Plaintiffs—Appellees,
versus
Joseph R. Biden, Jr.; Vivek H. Murthy; Xavier Becerra; Department of Health & Human Services; Anthony Fauci; Et al.,
Defendants—Appellants.
 
Oh yes....

We are using an old version of Bentley Microstation CAD software that does not work with Win11.

AutoCAD also does this alot. They release a new version ever year, with about 3-4 versions being on the same database. The first year of the new database are always messed up.
 
I had a fun one last year. Microsoft released an update for Windows Server, and, not sure who was drunk on release testing that one, but it stopped Hyper-V from working. No VM's would come on. So, rolling that one back. Of course they then re-release that update later, that doesn't break Hyper-V, but sweet Jesus, if it's an update for a server OS, you'd expect higher QC than this.
 
Software QA is largely a thing of the past. Now it's a rush to release to meet sales/marketing deadlines, and have users do the final testing, even for "stable" releases and not just "public betas," which have contributed to the lowering of standards, and expectations (thanks, Google!).

As a user, one can try to file bug reports, but to see them dismissed, or ignored in the tracker doesn't encourage keeping up the effort for all but the most dedicated (or annoyed with a particular bug).

Most software is mature, so a lot of new features aren't essential, making security the biggest reason to update (outside of a manfuacturer's push for new revenue generation though mandatory updates). Many users don't, but the risk/reward ratio still usually works in their favor, as long as aren't part of a targeted group, or engage in riskier behavior.
 
Back
Top