Worrisome State Of Software

[Alan Goldhammer]

This is where Apple has an advantage.  The system, for average users, is very locked down.  Advanced users can access any manner of things, but it's much harder for average users.  On Windows, it's just that little bit easier for average users to get wider access and therefore do some damage unless it's in a corporate environment where the MIS team has put in place proper policies / schema to lock things down.  And that, of course, is exactly why most companies lock things down.  It's not to protect against external hackers or the like - it's to protect against users, and rightly so.

And I do understand the sentiment and position of non-technical people.  Two cases from my own history (well, specifically my wife),

Your wife sounds a lot like mine!!!  She has an iPad which is supposed to be idiot proof but she uses it to watch Netflix shows.  All of the sudden it said she was out of memory which was strange.  Turns out she was downloading all the shows rather than streaming them and her iPad was full!!  She is on an overseas trip right now and I got a text from our cell phone provider that her phone had been charged for extra data use.  She sent me an email saying she didn't understand why this was the case as there was free WiFi everywhere.  I replied asking her to double check the settings (Android phone) to make sure that data roaming was turned off and of course it was not.  She turned it off and I trust that we won't see any further data charges.

[FabienP]

(...)
As I see it, the companies keep adding new features, push programmers to release the new software versions prematurely, instead of keeping things simple and reliable.
The frequent job turnovers don't help. Some of the programs (Windows, SAP, etc) are now so bloated that no one in those companies can possibly understand how the systems work, and they are just patching the programs and hoping it will work.
(...)

I have never heard of a project manager in software development that had the liberty to prioritise stability or performance over features, at least outside of some specialised areas (aviation, military, banking, etc). It's not (yet) a selling argument to a broad audience. Only when lacklustre performance really hits very high levels and attracts many negative comments, then yes, something gets done (see the emphasis on speed in the latest Lightroom development announcements for instance).

As for bloat, this can be explained by backwards compatibility needs and negative backlash as soon as a moderately used feature is removed or replaced by a "better" version (akin to the new import workflow in Lightroom 6.2). Couple this with new features that are regularly thrown in with no clear use case or with half-thought implementations and this will only raise pressure on maintainers, who will need to find ways to stay profitable.

I think that a change in how software companies deal with quality insurance is likely to make the life of users even more difficult. The list is from Microsoft, but this might apply to others as well:

• Cut costs by firing all compatibility testers and promote users to unpaid beta testers.
• Gather as much input passively from users using telemetry. Make it nearly impossible to opt out. Instead of only using this information for maintenance, sell it to make it even more useful now or in the future (since the terms of services can be changed for data that has already been collected).
• Make it mandatory to always run the latest build of your application so that no one can escape being a beta tester by using an older version.
• Make it really hard to download and install an older version.
• Progressively restrict the ability of users to perform certain actions, because it can lead to security problems. Say for instance installing unsigned drivers to use Argyll CMS with an I1 Pro.

So software companies are progressively working towards making their own life easier. The majority of users will put up with this because it is for their own good or it doesn't bother them at all for mainstream activities. Or because they have no alternatives.

Cheers,

Fabien

[LesPalenik]

I have never heard of a project manager in software development that had the liberty to prioritise stability or performance over features, at least outside of some specialised areas (aviation, military, banking, etc). It's not (yet) a selling argument to a broad audience. Only when lacklustre performance really hits very high levels and attracts many negative comments, then yes, something gets done (see the emphasis on speed in the latest Lightroom development announcements for instance).

As for bloat, this can be explained by backwards compatibility needs and negative backlash as soon as a moderately used feature is removed or replaced by a "better" version (akin to the new import workflow in Lightroom 6.2). Couple this with new features that are regularly thrown in with no clear use case or with half-thought implementations and this will only raise pressure on maintainers, who will need to find ways to stay profitable.

I think that a change in how software companies deal with quality insurance is likely to make the life of users even more difficult. The list is from Microsoft, but this might apply to others as well:

• Cut costs by firing all compatibility testers and promote users to unpaid beta testers.
• Gather as much input passively from users using telemetry. Make it nearly impossible to opt out. Instead of only using this information for maintenance, sell it to make it even more useful now or in the future (since the terms of services can be changed for data that has already been collected).
• Make it mandatory to always run the latest build of your application so that no one can escape being a beta tester by using an older version.
• Make it really hard to download and install an older version.
• Progressively restrict the ability of users to perform certain actions, because it can lead to security problems. Say for instance installing unsigned drivers to use Argyll CMS with an I1 Pro.

So software companies are progressively working towards making their own life easier. The majority of users will put up with this because it is for their own good or it doesn't bother them at all for mainstream activities. Or because they have no alternatives.

Cheers,

Fabien

That's a very good summary, Fabien

The sad thing is that bugs and poorly conceived and written software have nowadays become the norm.
It's one thing if an inexperienced user downloads some virus or malware, that is after all his fault or negligence, but if a national bank, eBay, Paypal or other major institution releases defective software to millions of users, that's simply inexcusable.

In practical terms, software bloats cause newer versions of computer programs to become perceptibly slower, use more memory, disk space or processing power. All these facts are known to most users and support technicians.

The bloats in software are due to some extent to keeping backward compatibility, but in many cases the old and inefficient section of code stay in place, simply because there is no budget to redesign and streamline the system. At first glance it seems easier to keep the old program modules and just interface it to some new lines of code. Quite often the maintenance programmers don't analyze and understand the legacy applications, and the IT managers don't have a clue about the program workings. Fortunately this doesn't happen in all companies, but in many more large corporations than the general public would believe.

[Alan Goldhammer]

In practical terms, software bloats cause newer versions of computer programs to become perceptibly slower, use more memory, disk space or processing power. All these facts are known to most users and support technicians.

Yes they are well known but totally irrelevant today.  Look at the difference in CPU speed & threads, low cost of very fast RAM, development of SSD technology to replace spinning platters, increased USB speed and devices.  At the beginning of the PC era look at what we had to live with and how software developers had to optimize progams.  I remember a wonderful text editing program that utilized the old WordStar keyboard shortcuts.  It was a DOS program but I cannot remember the name only that it was super fast and ran on PCs with minimal memory.

I was a post-doc at Cornell doing biophysical chemistry (1975-78).  We had a PDP-11 minicomputer in the lab that had to be booted up with toggle switches and paper tape each morning.  While we had both a Basic and Fortan compiler, the best way to get maximal performance was to have someone from the CS department come in and write machine language routines (nobody in the lab had that kind of background).

I spent probably $1300 building my current workstation and I have more power on my desktop than anyone would have imagined back in the 1970s when I was still actively using computers for research.  It's just mind boggling.

The bloats in software are due to some extent to keeping backward compatibility, but in many cases the old and inefficient section of code stay in place, simply because there is no budget to redesign and streamline the system. At first glance it seems easier to keep the old program modules and just interface it to some new lines of code. Quite often the maintenance programmers don't analyze and understand the legacy applications, and the IT managers don't have a clue about the program workings. Fortunately this doesn't happen in all companies, but in many more large corporations than the general public would believe.

Some of this is done because regulatory agencies demand it.  I spent the majority of my career in the pharmaceutical industry and there was a large amount of computer software validation required in the manufacturing area.  The industry used Windows XP OS for a lot of years because of the costs to validate new OS for mission critical computers.