Stephen's Personal Blog

Microsoft is offering a for-profit virus scanner. It is currently in beta testing now. This seems like an inherent conflict of interest similar to when stock ratings were closely tied to investment banks. Should Microsoft come out with a patch to fix a security hole or should it leave that problem only to be fixed for customers who have purchased the additional virus scanner? By purchasing this product you are provide positive feedback for Microsoft to back insecure products. With the Internet connecting many of the computers. Compromised or unsecure computers can slow down everyone else. And Windows is the most popular type of computer.

There should never be a monetary inducement for a company to compromise it’s ethics or social responsibilities.

I had a dream last night in which I rehashed an old issue that bugged me about my previous company. The smallest hard drive you can buy is massively oversized in a corporate environment. If all the new computers have 80 gigs and only 15 are used, then those computers are wasting 65 gigs a piece. For 30 computers you have 1.95 terabytes. This is as much storage capacity as a $10,000-$15,000 server. My personal problem, slightly related, is drives. I hate drive letters. My computer has a lot of content, almost 1 terabyte and it’s spread over 5 hard drives. I could care less what drive something is on, but the operating system requires that I care so that I can find it.

I want a program that will treat all the drives ( not including removeables ) on my computer as one logical space. Then I remembered UE’s problem and thought why not make the logical space include other computers on the network? My goal is to abstract away the drive letter and path. It may not be possible to remove it, but I should be able to make it just another piece of metadata.

Linux/Unix has something like this called LVM. The OS is tricked into thinking many “drives” are one. So, it writes a file across them all. But if one of your drives goes out all your files on all drives are lost. Three hard drives triples the chances of a catastrophic failure. Besides, I don’t want to learn a new operating system to solve my “minor” problem.

This is what I see so far. The current file based storage model is wrong. Filing something is very important and useful, but it is not the end all be all. It is another piece of metadata. There is also, whether the file is program or data. Kind of data; music, Word file, configuration file, saved game, etc. I want to drop a file into a box and open it without picking the file drawer.

Hopefully, someone else has already solved my problem.

Update 1: I was thinking about this this weekend. It seems I want a database instead of a filesystem. Every piece of ancillary data about a file should be in there. For example, mp3s of songs should have the lyrics, Word docs should have the plain text version, contacts should have all the contacts listed, etc. The would greatly enhance the searching aspect of so many files.

Version control (VC) is a basic tool of modern programming. It’s time for it to move out of programming and into mainstream use. In Word, you can write lots of pages; cutting and pasting, typing, etc. Press the undo button to get back to something or undo a mistake. But, what if the mistake you want to undo was a few days or weeks ago? Or what if the mistake was saving something else on top of the good file? VC provides you with an undo button for these situations.

The basic process works like this. You open a file and make your changes. When you close it you “checkin” the file and make a little note about what changed. Then something drastic happens like you accidentally copy the wrong file or you delete your file and it seems like there’s no way to get back it. Now, VC becomes very important. Open a VC program that lists all the times you checked in your file with the little notes. Select the one you want and “checkout” the file.

There are many other useful features of VC. For example, you can look at a list of files and see if one has been changed, but not checked in. You can keep several copies of a file up to date with one another. One of my favorite tricks at UE was writing code on any one of 6 different computers and at night they would get the latest version. Using Diff (short for difference) you can compare two files to see what changes there are. Also, it provides a backup in case of disaster. The Repository is where the many versions of a file are kept. Copy this one file or directory onto CD or another computer and you’ll have a cheap backup.

I expect to see some form of VC in the next version of Windows after Vista. Or possibly in a service pack. But, why wait? There are stable free programs available today that work extremely well. My favorite programs are Subversion, which acts like an engine, and Tortoise SVN, which lets you work with the files by clicking on them.