Windows 10?

For what we are paying for that software, it should have been written in Assembler Language. Looks pretty lazy, the bulky slow code they have right now.
 
It only seems like a fantasy world because the concept of well-tested-at-release, robust software has been so completely eradicated. But Pepperidge Farm remembers.

I cut my professional teeth on a "legacy" software running on *nix. In 2003 I spearheaded my first install as lead. It was for a small company who had spent shit on hardware, maybe 8 grand, so it was not top of the line.
They had about 1.5 million records to convert from their old system. Getting the datafiles in order took a long time, but getting the data inserted into that "legacy" system took about 90 seconds.

Today I'm working on whiz bang new wow software that uses your choice of Java application servers and database systems and multithreads and all sorts of totally unnecessary shit. And here we are 12 years later, meaning today's hardware is 8 times more powerful.

And this new system, if it contains hundreds of thousands of dollars of top shelf hardware, if you drive it until smoke is coming out of it, might be able to insert 10,000 records per hour.

It's not just Microsoft to blame. Those of us living this know that it's a dark time in the world of computing.

The main reason I use Linux is because so far it has managed not to succumb to any of that. My home machine is 2006 era tech. If I try to stream Netflix and run Wine at the same time then funny stuff might happen, but upgrading to 5 year old hardware will fix that. Hardware is ao good these days that they actually teach CS students to waste resources (I have the textbook to prove it). Linux doesn't brook that nonsense.
 
For what we are paying for that software, it should have been written in Assembler Language. Looks pretty lazy, the bulky slow code they have right now.

Faster hardware does tend to minimize and eliminate the software language speed relevance and performance differences.
 
Faster hardware does tend to minimize and eliminate the software language speed relevance and performance differences.
Faster hardware should be no excuse for shoddy programming. Imagine how much more efficient, both memory wise and processor time required if they actually took the time to do it right.
 
Faster hardware should be no excuse for shoddy programming. Imagine how much more efficient, both memory wise and processor time required if they actually took the time to do it right.

Definitely.

Software hasn't become cheaper either. Price/quality definitely went down over time. These days when version 5 of something comes out a week later there's already talk of version 6. Why ? Was 5 not good ? Sometimes seems that way.

In many cases I'm more than happy to run old software with less functionality since a lot of those functionalities make the software bloated and nearly unworkable. So yeah. I'm all about technological advancements but I hardly see any real advancements beyond version numbers and lots of fancy options you mostly don't need.
 
Faster hardware should be no excuse for shoddy programming. Imagine how much more efficient, both memory wise and processor time required if they actually took the time to do it right.
It always has been. 20% of the code most usually does 80% of the work.

OO just speeds up the new versions roll outs.

As an MS software director said ~20 years ago, if they're going to sell 100 million copies regardless, how good does it really have to be?

Good, fast, cheap. Pick two.
 
One class I had in Pascal, they had us writing a bubble sort. I took the time to write mine with two options, one was using a Pascal routine to do the sort and the second option was to implement an inline machine code routine to do the same thing. The screen was filled with random characters in the video ram and the routines then did the sort where you could see it happening right before your eyes on the screen. The professor was very impressed at the difference in speed between the high level language versus the low level language. You could see the sort taking place with the Pascal routine, but with the machine code routine, the sort was nearly instantaneous.
 
Last edited:
free-linux-wallpaper-widescreen-1920x1200.png


More will come.
 
One class I had in Pascal, they had us writing a bubble sort. I took the time to write mine with two options, one was using a Pascal routine to do the sort and the second option was to implement an inline machine code routine to do the same thing. The screen was filled with random characters in the video ram and the routines then did the sort where you could see it happening right before your eyes on the screen. The professor was very impressed at the difference in speed between the high level language versus the low level language. You could see the sort taking place with the Pascal routine, but with the machine code routine, the sort was nearly instantaneous.

Try it in C. Report the execution speed difference between C and assembler. Algorithm selection is key. Bubble sort sucks.

Hardware is cheap and getting cheaper. People are expensive and getting more expensive.
 
Kolibri is an unusual OS. I'm not sure if it will run any programs from either Windows or Linux, I think it only runs it's own apps. I wish it was more functional, but my understanding is everything is written in Assembly, and will say it does bootup very fast.

http://kolibrios.org/en/

 
One class I had in Pascal, they had us writing a bubble sort. I took the time to write mine with two options, one was using a Pascal routine to do the sort and the second option was to implement an inline machine code routine to do the same thing. The screen was filled with random characters in the video ram and the routines then did the sort where you could see it happening right before your eyes on the screen. The professor was very impressed at the difference in speed between the high level language versus the low level language. You could see the sort taking place with the Pascal routine, but with the machine code routine, the sort was nearly instantaneous.

I did something similar in a VB course. I had a project to pick lotto numbers based on some rules and got some extra credit for implementing it in C also.
The VB version took some ridiculous amount of time like 3 minutes per run. The C version didn't look like it took any time at all.
In fact it was a perfect learning experience for a first year student because the C version immediately uncovered the inherent problems with random number generation, because after the fifth run numbers started repeating.
In the fancy GUI version, I would never even have noticed that.

Fast forward 20 years and it seems like 75% of what I do is troubleshoot why the process is taking so long, and when we get reports of genuine business logic problems they all take a back seat to the lock ups.
 
Kolibri is an unusual OS. I'm not sure if it will run any programs from either Windows or Linux, I think it only runs it's own apps. I wish it was more functional, but my understanding is everything is written in Assembly, and will say it does bootup very fast.

http://kolibrios.org/en/



Can't complain about w7 from SSD's either. 3-4 seconds. Of course this will get slower over time but then it's time for a fresh install.
 
Can't complain about w7 from SSD's either. 3-4 seconds. Of course this will get slower over time but then it's time for a fresh install.

If you have to reboot Linux, at all, ever, it's a hardware problem.
If you have to reinstall Linux, at all, ever, it's a hardware problem.
You may wish to upgrade when your hardware problem reinstall makes it convenient.
 
Back
Top