Roy Schestowitz wrote:
> Dana rightly asks: Where are all the proprietary hits?
>
> ,----[ Quote ]
> | People are fond of deprecating open source for not producing a few
> | zillion jackpots. The problem with that line of thinking is that it
> | presupposes there have been a lot of proprietary hits. There haven't, as
> | the link above < http://tinyurl.com/krt4b > notes. There aren't that
> | many $1 billion-plus software companies.
> |
> | [...]
> |
> | It's just a phase the kids are going through, though. They'll
> | eventually learn to open everything. Everyone grows up at some point.
> `----
>
> http://asay.blogspot.com/2006/09/dana-rightly-asks-where-are-all.html
I think it was inevitable that the types of packages and number of packages
within each type would narrow down to just a few well trusted applications
and utilities. In the end there is a job to be done and to do that you want
applications that will allow you to get on with it by the easiest route.
So now we have a situation where it is unlikely that The Killer Mega Selling
application will appear again, I don't think that is because we have
already got everything that a computer is capable of. Much more because
application development can be a very costly business and users are less
willing to pay the value of that development.
If you consider that at one time it was taking an average of three years to
write an orriginal game program, let's say you have a team of 50 in design,
graphics, story, modeling, maybe actors too. That means you have to pay up
front the wages of 50 skilled people for up to 3 years, and you have no
idea if your product will be successful. Remember that in the Games Gold
rush they were many major successes, but they were many many more failures.
The same is true for major applications. Which is why it is better to build
on top of something that already exists and is in use by a lot of people,
such as addon packages. Then with a shorter development cycle and hopefully
with a good estimate of how many users will be interested in your product,
you stand a chance of making a living.
I personally would like to see the commercial side of Linux picking up. Yes
I know, everyone wants it all for free, but it isn't free and never was, it
always cost someone in time and effort. Until they are rewards, or at least
a decent living to be made, there will be a lot of potential linux
programmers, experts in various fields of programming and diagnostics, who
wont come into our corner of the room. We have good programmers and of
cause the teams from the likes of IBM, HP, SunSystem etc, but we need more.
There is another area of cause where the propriatary software can rear it's
head again. There has been talk of it being time to abandon the IBM-PC
design. If you loaded one of the *DOS OS's, then ran your old asm or C
hardware control software, it would still work. But it may be time to
abandon that compatibility and head in a new direction.
That isn't meant as a change for change sake. It is because the current bus
technology is growing messy, trying to cover too many possibilities at
once. A brilliant design of cause, but no longer valid for the wide bus
machines. For example, it is now quite possible to have all IO addressing
in a single space with most instructions plus possibly some data in single
instructions, such as a comms port, remember how you would stream about
eight control codes to it to set it up, then have to link in a dma
channelfor the ring buffers (if you wanted any sort of decent speed), Well
such things wouldn't be necessary if the machine's development was started
here and now. Probably we would design in an intelligent wide bus comms
chip. Just about every interface would be in for a major change. Gads, what
developer wouldn't make majort changes to the clock system in the PC.
I am not saying the IBM-PC is a bad design, only that with what we know now,
the design could be much better.
|
|
