Human Readable   

 

     
   
     

Slackware 10.0—Day 13

© Copyright Darrell Anderson.

After my time-consuming VectorLinux (VL) experience, I decided next to try Slackware 10.0. VL is based upon Slackware, although always a previous version (weird—why not just fork and move on?). While visiting in the city I downloaded Slackware 10.0 stable, but I also downloaded Slackware-current. Slackware-current is notoriously stable, but simply not labeled as a milestone edition.

Some people will argue that going from VL to Slackware is like jumping from the frying pan into the fire. Possibly so. However, one reason I initially tried VL is I want a minimal distro that allows me to grow and expand as I choose and not as a vendor chooses. I want to avoid vendor lock-in. Slackware and VL fit that description, although VL restricted the way I wanted to install the software. I already have tried several of the newbie distros and I do not like the way they attempt to control me and my computer. This is a personal decision. Most people migrating probably should stick with the newbie distros.

I’d like to address a point here. All but a few distros are available for free if one has the bandwidth to download complete ISO images. CDs also are available from many people for little more than the cost of the disk. This is a wonderful deal. I have not yet paid for either VL or Slackware and do not intend to pay anything unless or until I decide that distro is going to remain on my computer and become my workhorse. I do believe in supporting distro developers and I have done so in the past, but I also believe that one should be able to try before providing that support. There must be a mutually benefitting and reciprocating relationship. This is a huge problem with proprietary software. One must pay first and then pull teeth to obtain a refund if one does not want the software. I never liked the cripple-ware marketing approach either. Crippled software cripples the evaluation. The way current distro developers offer their products seems to me the proper way to go. If the VL developer had provided a download version packaged with KDE and supported multiple partitions, I might still be tinkering with that distro. And yes, eventually I would support that developer if the overall package won me over. That did not happen.

Before starting my adventure with Slackware I perused the many text and readme files. I decided to print the installation How-To. I was using Mandrake 9.2 and had long ago configured CUPS for my HP LaserJet 4200. I have a duplexing tray installed. I started to print but stopped to check the duplex option. Nope, not set. Then I continued. As soon as the first page printed I knew something was awry. I was using Kate. The text was in huge characters and the job was trying to print outside the page. Some kind of configuration problem, but where? I wanted to move forward and not chase another wild goose.

I then looked in the Slackware CD for some preformatted PDF files. None. I stopped, rebooted into Windows, loaded the text file into Word, added page numbers in the footer, and printed. Done.

I was still smarting from my hours with VL. I admit I was running a short fuse at this point. Sure, had I taken several minutes to investigate my Kate configuration I would have solved my problem. However, I’m not alone with such shortcomings. Even experienced ’nix people sometimes can’t get some of this stuff to work out-of-the-box!

My frustrations continued as I tried to install Slackware, although initially to no fault of the Slackware maintainer or software.

Using the Mandrake 9.2 DiskDrake utility, I created three additional ext3 partitions. When I rebooted, I realized DiskDrake had changed my MS-DOS partition to hidden, and had changed my FAT32 partition to hidden FAT16. Why are the Mandrake programmers messing with those partitions when they update the partition table?

I already had separate swap, /boot, /home, /tmp, /opt, and /usr/local partitions on my hard drive. After perusing the Slackware How-To, and with separate 2 MB /opt and /usr/local partitions, I thought 256 MB /var, 2048 MB /usr, and 256 MB root partitions would be sufficient.

Like the VL installation, the Slackware installation routine recognized my elaborate partitioning scheme. Unlike VL, Slackware allowed me to manually configure my fstab using those partitions. I had already known about this difference, but nonetheless—great!

When I proceeded with the installation, I knew immediately that the Slackware /usr partition was too small because I began receiving error messages when installing the GNOME software. Either I misread the Slackware How-To or the How-To is misleading.

I returned to my only working GNU/Linux environment, Mandrake 9.2. I resized the new Slackware /usr partition. Once again DiskDrake bit my backside. This time, DiskDrake simply erased the partition type information from my MS-DOS partition! I’ve been down this road before when I was learning about partitions and I expected that all of my data was still intact. I booted with an MS-DOS floppy disk, then started the PowerQuest Partition Table Editor. I changed the partition to type 06H (MS-DOS). I also once again had to change my FAT32 partition back to FAT32. Yes, I maintain backups and I could have restored everything if the partitions had been destroyed, but why am I required to undo this nonsense?

I rebooted and verified all of my boot selections. No problems, everything was working as expected.

I rebooted with the Slackware CD. Once again I received the same error messages when installing the GNOME software. I knew right away what had happened. The DiskDrake utility resized the partition, but when I reformatted the partitions using DiskDrake, apparently the format routine did not change the file system size to match the partition size.

I rebooted into Mandrake 9.2, opened a command line window and executed the e2fsck and resize2fs commands. Sure enough, DiskDrake had failed to resize the file size. Aargh!

I rebooted once again with the Slackware CD.

Unfortunately for me, I experienced fat finger syndrome and selected the incorrect partition for my /home partition. I tried undoing my selection mistake. No go. Then I realized the Slackware installation routine provides no “Back” or “Undo” feature. I did not know what to do, so I pressed Ctrl-Alt-Del and started over. Aargh!

I tried again, but as my luck of the draw would go, I made another typing error. This time I pressed the Esc key, hoping that action would serve as a “Back” button. Nope, the action was a Forward/OK button! I performed another three-finger salute. Aargh! Enough for one day! I shut down the computer and spent the remainder of my evening reading.

At this point experienced people might think my story somewhat humorous. Because there is no long-term damage, I suspect that one day I’ll laugh too! Not today, however. Yet, I’d like developers to consider one point: visit any discussion forum and realize that this kind of thing is happening far too often. What kind of message does that send to people who are considering migrating to GNU/Linux?

I find hard to believe that GNU/Linux has existed for the past decade and a half, in parallel with both Windows and Macintoshes, and many GNU/Linux people still don’t “get it” with respect to usability. Why do many GNU/Linux developers and programmers consistently insist upon doing things the hard way? Is this some kind of unwritten red badge of courage?

Finis.

Table of Contents