Jeff said:
Have you actually asked users what they want? ...
Yes. That used to be part of my job especially in the med biz. It is
still my job in self-employment. One of the questions on my list for the
first architecture meeting is about error indicators.
... I was thinking very
much along your line of thinking, until I decided to actually ask.
What I got was what I described. Most users don't want to see
anything if the program barfs. They just want it to work and consider
an error message a sign of failure. ...
Because they usually are
... Incidentally, the more detailed
the error message and remedial instructions, the bigger the failure.
While this is admittedly overly simplified thinking, it's one of the
major reasons that both MS and Apple have dumbed down their error
messages. Elaborate error messages scare the hell out of the GUM
(great unwashed masses).
The effect of such dumbing down is often a backfiring. Grandson tries to
fix aunt Brunhilde's computer, sees numerous meaningless error messages,
says to her that the software (or the whole computer) is junk and she
should return it for a refund, then get one that works. No joke, this
has happened in our neighborhood. The whole computer was returned and
they bought a new one, which successfully connected them to the Internet
(that's all they wanted).
Some programs actually do that. It doesn't seem to work, as nobody
seems to be interested in helping the software authors fix what is
perceived to be *THEIR* problems. Even "click here to send debugging
mumbo-jumbo to the developers" tends to get minimal responses. What
seems to work is checking a box during installation that indicates
that you approve of sending debugging junk to the developers in order
to improve the quality of their product. Expecting users to do
anything helpful is wishful thinking.
I think we may live in different worlds. I (and pretty much everyone I
know) tend to call if some important or expensive piece of software
doesn't work. The service guys at the other end of the phone line often
become rather desperate when error messages are bland and
non-informative. Because then they can't diagnose. I've had them say
things like "Wow, we've never seen that one before" or exclaiming "I
can't believe this is happening!".
Not really when you consider the alternative of having the entire
machine roll over and die at the first sign of a minor error. I would
prefer it would limp along as best it can, even with a big chunk of
RAm permanently allocated to a dead process, than a panic and reboot.
A good analogy would be having a blow out in one tire on a vehicle. I
would much prefer to have the car continue to roll on 3 wheels than to
have the car quit just because one tire is gone.
Except you may get into a major crash if that blown 4th tire shreds up
and locks up some stuff that was supposed to remain moving.
Even when (and if) Oracle eventually fixes Java? It's not like Java
is going to disappear overnight.
Fixing ... trusting a fix, two very different things
If a web site wants Java I just move on.
What I usually do in such debates is to calculate the cost of
providing a similar service using rented bandwidth and rack space. My
guess(tm) is that depending on usage, $50 will be rather cheap for the
necessary bandwidth. Based on the numbers from:
<
http://www.webex.com/pdf/wp_bandwidth.pdf>
The required bandwidth can be as high a 500Kbits/sec per session,
times two if you run full duplex video. So, what does 1Mbits/sec cost
these days? I dunno and I'm not going to dig out the cost of an
OC-192 right now (because I'm late for dinner). Maybe later.
Again, then how does Skype do it for $4.99? AFAIK that is the
non-ad-supported plan where this payment becomes the only revenue source
for them. After a year it pop to maybe arlund $10 or so, still not $50.
I'm jealous. My first guess on things that I'm intimately familiar
with is usually right. My first guess on things I know little about
is usually wrong. I can compensate for my lack of experience by
extensive reading, Googling, and simulations, all of which take time.
The challenge is to bring whatever your first guess is over to the folks
at the other location. Without live desktop sharing that can be quite
cumbersome when dealing with complicated matter (like a switcher layout).
Incoming video probably works just fine. So does audio both ways.
However, I have my doubts about the outgoing video at 128Kbits/sec.
That seems awful low outgoing bandwidth. Is that all your ISP offers
or are you being throttled due to a crappy phone line?
I'll have to call them again on this but the line to the next box is
rather long. But I do not need to broadcast a ballgame in HD, the
display of my desktop (but mostly that of others who need my advice) is
what's needed. For example, I have recently coached a team of mechanical
engineers to use an oscilloscope for the first time in their life,
diagnosing a fairly tricky timing issue across roughly 1000 miles. It
worked, and fast.
On the Skype problems, I previously posted instructions on how to rip
the monster out of the system and registry. Then reinstall from
scratch and it should (hopefully) be better.
I've done that. It no work :-(
Well, if you're desperate, you could just patch the audio from a
conference line into your computers audio system, mix well, and serve
to the rest of the participants. There are better ways to do this,
but for those that can only get a POTS or cellular voice connection,
it should be adequate.
Sure, if necessary I will do that.
I think you might find the politics interesting. Microsoft bought
Skype for $8.5 bazillion dollars without a clue as to how to sell the
service or generate a profit. I don't want to get into detail, but
VoIP profits are generated in the tiny fraction of a penny difference
between wholesale bandwidth costs and per minute charges. There are
an awfully large number of minutes involved, so the numbers can
potentially become huge. However, small mistakes can also expand into
huge dollar losses. MicroSloth hasn't learned this yet, and may be
trying to use it's traditional method of (almost) giving away the
product in the hope that the competitors will roll go broke competing
with a nearly free service. We may soon see how well that works after
MS decides to stop diluting the numbers by mixing the Skype financials
in as part of their Entertainment and Devices division numbers.
Hmm, that would explain it. But not quite, because there are other
services that MS doesn't own that offer a simlar price range. For
example, Fuzebox was pointed out by one participant in this thread and
their basci service starts at $15/mo (if you don't need mobile hosting).
I forgot about that. That's another time burner for those without
autofocus cameras. About half the time, the product is out of focus.
We do MUCH better using video from a digital camera or DSLR both of
which have autofocus.
Actually. CAD software doesn't need a DSLR camera
I got an Olympus PEN camera plus adapter so I can use my classic Rokkor
lenses. Works great but no auto-focus. Doing it by hand is often better
anyhow because the autofocus can't possibly know when what I want in
focus. What irks me is that such modern cameras don't have a high-pass
focus helper. It would be so easy but ...
Yep. That's one of the benefits of doing meetings in real time. I'm
just questioning whether it's worth the hassle as compared to using
email and Dropbox.
Dropbox is nice but too slow for that. Went through this very exercise
yesterday around 10:30am. The file was 12MB so we had to discuss the
next topic and then go back once it had made it into the PDF reader of
all the others.
Good, supported, cheap. Pick any two.
Doesn't have to be cheap. $50/mo is ok, but should allow switching
saddles with the host account holder because of frequent travel. We
can't buy four accounts at a grand total of $200/mo just because of
that. At least not in this start-up phase.