[Date Index] [Thread Index] [Author Index]

Re: Re: Mathematica skill level snippet(s)

• To: mathgroup at smc.vnet.net
• Subject: [mg105016] Re: [mg104943] Re: Mathematica skill level snippet(s)
• From: John Fultz <jfultz at wolfram.com>
• Date: Tue, 17 Nov 2009 05:18:36 -0500 (EST)
• Reply-to: jfultz at wolfram.com

On Sat, 14 Nov 2009 01:56:14 -0500 (EST), M.Roellig wrote:
>>
>>
>> I did start to write a solution to do this, but there's a bit of
>> HoldPattern magic and such that needs to be done and I was too busy to
>> spend time getting it to work (I did have something kind of working,
>> actually, but then accidentally quit my session w/o saving, and gave up
>> because of the ten million other things on my to-do list).
>>
> If there only was an autosave feature ;-)
>
> (Sorry, but I could not resist)
>
> Markus
>
> P.S. BTW I agree to those that find the documentation helpful. It takes
> some time to dig into it, but one finds surprising stuff. Of course it
> can always be better but compared to other software docs it's quite ok.
> However, it is not always easy to find the information that one may need
> at the moment.

I don't know of any auto-save feature which would have saved me from 
myself...i.e., bypassing both opportunities to save during the session and a
request to save at the end of a session (every auto-save feature I've seen 
throws away info after you intentionally decline to save a document).

FWIW, we do have a naively implemented auto-save feature (and have had for many, 
many years).  It's not turned on by default for good reason, as was pointed out 
by another response to this message.

I certainly understand why there is sometimes frustration about stuff not coming 
out of the Wolfram feature pipe fast enough.  I think that, sometimes, people 
leap to the conclusion that we're not listening, or are being stubborn.  On the 
contrary, as I think many people around here do understand, we take many of our 
cues for features from comments on this list.  I don't know of any software
product where as many developers are as plugged into the user community (largely 
through this forum) as ours is.  However, we try to make a practice of not 
shipping features until we can make them a fully integrated and highly usable 
part of the system.  That's sometimes very difficult and time-consuming.

There's evidence of why we should do this in the archives of this group. For 
example, after frequent complaints about copy/paste (particularly of code into 
posts on this forum), we completely revised the feature in version 6.  It was 
something like a full year after we shipped this feature before anybody 
noticed...presumably because everybody had been trained by experience to avoid 
it.  Negative reinforcement is really strong and hard to break.  I've had 
several experiences like that which cause me to realize that people have a very 
long memory about stuff that's broken.  Better to not ship it than to ship it 
broken, is the conclusion I've reached.

So, whatever your pet feature is, if enough people on this forum have talked it 
up, we're probably working on it or at least have plans for it.  In some cases, 
we're even designing features based on MathGroup posts for which no feature was 
ever explicitly proposed (when the same shortcoming comes up again and again, 
maybe we can design a feature to overcome it).  I, for one, rarely respond to 
these things, though, because I hate making promises I don't know if I can keep 
(e.g., about when such a feature can ship), and because I'm pretty much the
wrongest person in the company to do a marketing/spin-job response.

Sincerely,
 
John Fultz
jfultz at wolfram.com
User Interface Group
Wolfram Research, Inc.