Ask Klaus

Open to Possibilities

Hi Klaus,

I'm an old PC/Windows guru from the days when PCs often didn't have CMOS or Real Time Clocks and adapted to them. In Linux Magazine 150 (my first issue), Neil was asking about Linux on a notebook with CMOS issues, and it caught my eye.

First, I've never known CMOS storage to fail, but after 3 or more years the backup batteries can easily fail. In a desktop, you just replace what is typically a CR2032 cell. My own desktop will need just such a battery anytime; it's been going a little over 5 years!

A notebook may be different as the CMOS may get its backup power from the main battery, which in this case is dead. I searched on the notebook specs provided in your column, and a battery popped up for US$ 62.99 plus shipping (US$ 8.00).

However, further search complicates matters, so helping Neil means knowing where he is and more about the computer. I cannot find any information for an Acer TravelMate 3750 at acer.com.au (my location) or acer.com. I can find an Acer Aspire 3750, so that may be what the battery fits.

Yes, I could have been blunt and said Neil should buy a new battery and remove the CMOS issues. But that isn't me, and I myself have had devices with unattainable batteries. The truth, though, is that Neil will probably NEVER eliminate these problems without a battery to keep the CMOS settings. In the worst case, he could break into the notebook to find either a backup battery, or to arrange an alternate backup battery.

Now, you might be interested in why this is my first Linux Magazine issue. I stick with Windows until I have a need for something else. I am working on some prime number matters in which it appears that I need to learn Sage or Python to exceed the numerical limits (2^63 – 1) of what Windows seems to offer. With 149 free issues on DVD, Linux Magazine looked like it may offer a direction to pursue … .

Now I have to deal with the plethora of Linux distributions, many of which do not even install readily under Hyper-V on a Windows 8 Pro Samsung notebook. Then there are Solaris (I have a little background here), the BSD collection, and even maybe Python on Windows to consider. Starting with the mess of Linux variants, I just wish our IT industry would get rational.

Not unifying Unix was a huge mistake, a victory for mindless dogmatic passion over the needs of everyone for stable, standard, and reliable platforms. Oh, how I long to work on IBM's mid-range business systems again!

If you can suggest a rational solution to get my project moving again, go ahead.

David

Answer

Hello, David:

Weighing effort and gain, it seems to me that you are not up for a system change from proprietary to free software (or Windows to Linux) at all. You are just looking for a programming language that will eliminate the need to rely on the CPU's and operating system's limitations in the calculation of large numbers, so Python with its corresponding math libraries on Windows (!) may actually be the most effective way for you to go. Don't bother running Linux in a hypervisor for this. Python, Perl, or Java with the appropriate libraries may be able to solve the problem on your existing Windows platform, so you don't have to look for a new OS.

It may surprise you that I don't recommend trying Linux in every case, but because you seem to feel comfortable with the system you have used for many years, and there is a solution that gets the job done on this system, why should I talk you into something else? :-)

However, because you mentioned the often heard criticism about "too much diversity" in Unix/Linux systems, please let me comment on this. In my opinion, it is a good thing that an operating system with applications can be split into many flavors of client/server/embedded/multimedia systems, it is not a disadvantage to be capable of individual changes.

As you know, Linux dominates the market in smartphones, and even though each manufacturer tries to do things differently, it is still the first choice just because it is flexible enough to be changed according to the target hardware platform and alleged customer needs.

I don't think that a single look and feel is really a "must." Different people have different needs and may be able to work more efficiently with a system that is assembled and customized especially for them. Having a single, monolithic, and static operating system and apps that always look the same would, in my opinion, be a loss of possibilities and potentials in building new and innovative software.

You also did not mention the fact that Open Source is not aiming to be just adequate or equal to something else or to "compete" with proprietary systems. Open Source is rather a business and development model that helps inventors share their ideas and work together in a cooperative way.

So, you are thinking that solving a problem with large numbers can be accomplished by just learning a new programming language or using a platform optimized for the task, and you may be right, but why not let a community of mathematicians, programmers, and analysts assist and help you find a good solution, rather than doing everything by yourself?

For complicated problems, working together with others in open projects and learning from their methods to solve problems may greatly speed up learning and understanding. This idea is entirely unrelated to any specific operating system or programming language. So, my recommendation would be for you to describe your mathematical problem in a forum of computer/algebra scientists. They may come up with a possibly surprising answer that does not require you to switch to something you don't like.

About the battery issue: I agree with you that the obvious solution of replacing the battery because of a failing CMOS is surely better than a "workaround" software solution like syncing the clock on every reboot; however, the issue may also be avoiding breaking an otherwise perfectly running system – or at least one with an error that is known and can be handled by the user.