Why Linux over NetWare
Novell Cool Solutions: Feature
Digg This -
Posted: 13 Sep 2005
Question: My question is: why Linux over NetWare? I live in both worlds and I don't see a compelling answer to that.
Stomfi wrote back
You seem to be doing alright.
I would say the important answers are:
- OSS is obviously the future. Even Novell thinks that.
- You can get the best of both worlds with a NetWare stack on Linux.
- 300,000+ virtually free OSS software development community.
- OSS has the potential to deliver the long awaited all talking, all seeing, all reading "Model T" Human Interface, proposed just before MS cornered all the money.
- P2P WAN load sharing Grids, possibly using the Cell architecture.
- Change is now inevitable as the MS lid on progress has been kept firmly down for 20 years, and this is much too long in a developing technology, that was at the "steam powered moped stage" in 1985, and is still there.
The thing to remember when looking at developing technologies, are that the development of the parts follow defined 3 phase cycles. Each phase lasts 10 years, the first being research, the second, development, and the last sales. Each 10 years a new better part will replace the old one. So you can see from this that MS has been selling a part of the whole (in the vehicle analogy, the starter motor and key) for 20 years. (All the research and development in both hardware and software was done by others prior to their sales phase. Just look at an '85 SGI Indy if you want a comparison). And how long has 32-bits been in the starter motor?
Also remember that the parts are not the whole, and each part is a step, which once marketed can be ignored as a pointer to the future. NetWare probably fits this analysis, whereas Linux describes itself as a work in progress, so, although being marketed has still got one wheel in the development phase.
What do you think will replace Linux? My analysis is something along the lines of Plan 9. Can you get a NetWare stack for that?<LinuxNewbie@novell.com> wrote back:
- Novell, sadly, has been wrong before. I'm not saying they are in this case, but I think "obviously the future" is a bit of an overstatement. You might accurately say it's Novell's future since they're betting the farm on it, but it's not obviously everyone's. "Free" isn't a sustainable business model .
- Maybe. We're having 15 years of NLM optization being tossed out the window. Do we also lose the superior NetWare file system in place of the standard linux ones (ext2/3/reiser, etc) and lose the wonderful NetWare permissions model in favor of the crappy posix one? (I genuinely don't know the answer to those two questions, but if it's "yes" I have my doubts on how much progress that'll be).
- May be good, may be bad. If no one deigns to do what you need to have done, you have to do it yourself anyway. Don't mistake quantity for quality.
- Thanks for the belly laugh. OSS is good for lots of things, but originality and innovation (other than the model itself) certainly aren't present. Just look at the pieces of crap that are Gnome and KDE (Windows redone badly), among so many other things. In a lot of ways, OSS developers in general are like the Japanese manufacturers of the 1980's. Nothing original; they just took other peoples' ideas and made them better. Apple they aren't, and will never be.
- I can't even comment on that since I have no idea what that means.
- Huh? The MS lid on progress? Change is now inevitable?? Is this the OS dialectic?
Answer: For SUSE
I suppose I better answer this by putting my personal secret weapon for predicting the future into the public domain.
One of the tools one must use when predicting the future, is the past, as much of it as is relevant or one can mentally grasp.
This is a model of my secret weapon:
Click for larger image view
You will have to modify this chart to fill in all the bits I've left out, and the dates I've got wrong.
It also has a Z axis, which covers the continued marketing of each significant development step. This hopefully puts money back into the future trend, but in some cases especially where an old idea is marketed to the uninformed masses, or is so useful nobody wants to give it up, can take it out.
I also thought you might be interested in the predicted architecture circa 1982 of the robust computing model.
This is a simple 4 layer model.
Public Information Centre (PIC) Layer
Work Station (WS) Layer
P2P world wide networks of workstation class computers
Human Interface Devices (HID) Layer
Visual Display Devices
Writing & Text Input Devices
2 & 3D Visual Input Devices
Audio Input and Output Devices
Movement Control Devices
Consumer Robotic Devices
High Speed Interconnects between:
PIC and WS
Elements of WS (CPU channel, Storage Channel)
WS and HID
You can see from the model that the WS layer can be physically at some distance from the HID layer.
The key elements are a universal Operating System, and high speed networks. Will NetWare fit the bill or will the addition of Linux (UNIX) help? When thinking about this, remember that UNIX came from the laboratories of the world's leading communications organization of the day. Plan 9 also comes from there, written by some of the same developers as a logical replacement.
Another productive way of discovering the next best thing, is to look at what the majority of development laboratories, software and hardware, are using as their work horse at any given time.
My secret weapon chart can be used to predict in other subject domains. It doesn't rely on luck, luck being in this case, inattention to detail.
This may not specifically answer your question, but it will tell you how I got my answer.
This Newbie answer has covered:
Novell Cool Solutions (corporate web communities) are produced by WebWise Solutions. www.webwiseone.com