At TechEd 94, Myrvhold - (Ex MS Research Chief), made the statement that
Software is like Gas - It expands to the size of the container...
laughed, of course, but its not so funny. I just had to take my OS and
browser version backward to Win95/OSR2 and IE 3.0.
So what? Well go and see an MS website page and you'll find that they
support an IE 3 CSS Style sheet - Nice and polite, very considerate,
Now go a KB support page. If you use the SHift key and scroll down to
select the text, say on Q129635 (or any page), it doesn't let you select
and copy - I don't think it did before either.
BUT, the big deal is this... The numbers on the pages (Step 1, 2, 3, etc.
are really there! The CRLF, in the Applies to header, don't work in IE
- Its just one continuous line.
Here's the impact - my reading of it.
A lot of extra code is being added by developers at all stages (from MS
you and I) to parse/filter and avoid the additional ZERO/NULL syndrome.
BLOATWARE is self-perpetuating.
In VB, this started with the VB4 version A release (I think they refer to
it as a Syncronizing release) You have to recall in Bruce McKinney's
book, that he knew about VB5 on the books, but not the VB 4.0a version
(4.0a does binary 0 compare by default) 4.0 didn't.
If you look at a lot of Parsing routines, they tell you to include the
delimiter at the beginning and end of the target line... to make it
That's how bloatware is perpetuated: We have to add stuff to get around
As seen in this little "Back to Test" - WIn95/IE3.0, I wonder if whole
pages aren't spun off this way.
My Registry files are 25 percent of the size they were yesterday! All
functionality and applications identical, EXCEPT no IE 5.0, no VB5.0.
Go to your kitchen and take out a knife. then go to your workshop
take out a hammer. Now go to your attic and dig out a plain-jane Ma Bell
telephone. Notice something?
They all still work today! They were built to serve us - the user. Not
advertiser, not the industry who made them.
I think its time for a new paradigm, a new computing window. We aren't
headed that way, with these current players. Not one of them, even MS is
really able to do it.
Our software industry is becoming like the self-perpetuating creeping
increase of Bureaucracy in either Government or IBM or GE. And so are the
products! They're tweaking old methods to make them look "Break-through"
Smoke and mirrors without underlying structural change.
I'm CCing our buddy Bill, because he might be able to spearhead a new
shift, but if not, would you like to see if we could find a few Paradigm
Makers? You never know, if Bill Gates has reached his plateau and lost
sense of challenge, maybe there's room for new, practical, really useful,
non-fad, low-fat software and hardware architectures.
I absolutely agree. Remember when you could boot a complete operating system
from a floppy disk (the old kind that could barely hold a grocery list, not
the new 1.4MB floppies or the even newer 100MB floppies). And when 16K RAM and
30MB of hard disk space was luxurious? My watch has more computing power than
NASA did when they went to the moon.
I think some of the problem is haste. Companies rush to shovel products out
the door with all kinds of bells and whistles so you feel you need to buy the
next version of the product. The reason you have had the same knives, hammers,
and phones for years is that noone can convince you your knife will be obsolete
next year. (You're a whimp if you don't get the latest knife with auto-cleanser
and spell checker!)
Just a small voice from a non-programmer with a habit of buying program languages and associated expensive books. I have read your comments [in How to Become an Expert Programmer] with interest. I'm not a programmer and strongly doubt that I have the stomach to be one. But other than having an immense desire and a lot of time to learn programming, and perhaps some professional assistance, I don't think that training to become an expert is made particularly easy from the start.
Take the "learn X in 21 days" series, these are expensive and poorly developed, though from the number of pages you would think that they would be comprehensive (just take a look and see how many of these books spend most of the pages dealing with explaining the UI rather than how to do programming).
I've just started with the VB6 learning edition and was impressed with the introductory CBT but within a very short while found myself asking some serious questions - database handling for one, very little was dealt on the subject and your left to wade through masses of verbals in the online help.
What I would like to see is a series of training books called "No Frills Training on X" that just stuck to the facts without a lot of pleasantries and unimportant stuff like half a book on the UI. Then maybe I could retain my enthusiasm long enough to learn one of these languages. (Ps I'm not a computer nurd like I'm afraid most programmers seem to be, so assimilating manuals is a slow process)
[I have actually tried to sell this idea to my publisher. A series of small, perhaps 100 page, books on very specific topics: how to print, how to use a simple database, etc. Unfortunately these books would be fairly expensive for their length so my publisher feels they would not sell. I think you need at least a couple hundred pages these days. -- Rod]
I agree that sometimes the hunt for the great book is a hard journey but,
if for nothing else, the learn xx in 21 days books do allow future
programmers a glimpse at what they may need to know without scaring them to
death which is what the high tech book could easily do.
I will admit that I have bought a few to see if this is what I am really
I absolutely agree. Remember when you could boot a complete operating
system from a floppy disk (the old kind that could barely hold a grocery
list, not the new 1.4MB floppies or the even newer 100MB floppies). And when
16K RAM and 30MB of hard disk space was luxurious? My watch has more
computing power than NASA did when they went to the moon.
I think some of the problem is haste. Companies rush to shovel products out
the door with all kinds of bells and whistles so you feel you need to buy
the next version of the product. The reason you have had the same knives,
hammers, and phones for years is that noone can convince you your knife will
be obsolete next year. (You're a whimp if you don't get the latest knife
with auto-cleanser and spell checker!)
I think another face of it is a syndrome I call "throwing good time after
As development continues and features are added, every program, every OS,
every architecture eventually reaches a critical mass where certain core
assumptions of the design aren't valid anymore. They are too limited or
constraining or just plain wrong. At that point, the developer/engineer has
a tough choice to make:
- He can pile workaround on top of workaround to make the old setup hobble
along as best it can, or
- He can bite the bullet and re-do the thing from scratch.
Unfortunately, very few individuals are willing to go with option 2 or even
consider it. The idea of re-doing old work is simply unthinkable to them --
it amounts to admitting that you are fallible and made a mistake(!) So to
preserve the old ego, they'd rather spend months or years trying to turn a
Dodge Dart into a Formula-1 racer, than to just go back and build the racer
in the first place (which would probably take far less time and trouble).
Caution point: while option 2 needs to be used a lot more often than it is,
we should be careful about going overboard with it. The opposite syndrome
(which I call "Stevejobsitis") can be just as bad.