ANN: Rosetta Booster Beta 1.0.1a (reference [NSTask] -launch return)
ANN: Rosetta Booster Beta 1.0.1a (reference [NSTask] -launch return)
- Subject: ANN: Rosetta Booster Beta 1.0.1a (reference [NSTask] -launch return)
- From: M Pulis <email@hidden>
- Date: Mon, 15 Jun 2009 08:31:39 -0700
Group,
In a failed attempt of the group to provide a solution for a recent
poster, I became intrigued with the subjective analysis that the OP
represented as facts responsible for rejecting (with extreme
prejudice) my proposal in particular. Others may have received off-
list personal rejections; I can only speak of mine.
My purpose in sharing the story here is to first acknowledge those of
us that took time to attempt to assist / educate the OP.
I felt, also, guilty, for having provoked the OP by labeling the OP's
concerns as "Princess and the Pea" while the OP referenced Dr.
Tannenbaum. With an epiphany I then realized that not only did we
read different books.... but, more importantly, that was absolutely
unfair of me to belittle the OP's issues without objective data.
To humbly apologize for my professional transgression and show due
respect for those passing subjective analysis as fact by presenting a
virtual olive branch, I immediately challenged the OP to create a
business case for Apple DTS changing how Rosetta works. Then, awash
with the guilt of assigning such an arduous task I asked myself what
would it take to brew such a case? And how much brew would it take to
even believe such a case?
Knowing that objective data would be required, and, at least in my
shop, to even approve a DTS expense, or explain to a client on
invoice, I took the suggestion by Mr. Guerin and others to prime the
pump. I implemented an exit(0) app. And tested it as a log in items
app to acquire the relevant data. In the discussion below, I have
included significant specific references for further study by
interested parties.
Second, to expose Rosetta for the resource hog it must be to have
created such scientifically objective responses as "a really really
bad idea". And "Nooooooo thanks!". I know that "double really", like
cooties and capital letters, is MUCH (see?) more serious than just
one "really", but, just how much more is "double really" really bad?
Maybe it is really bad squared.... that would really be bad. And how
bad was it to then force the keyboard into auto - repeating the
letter "o"! Not only a "double really bad" rating but also a seven
"o" No rating? Wow! I feel like an awards show... "you hate me, you
really hate me!" But I digress..... :-)
I set out to answer these questions: just how bad is really bad
squared? How bad is No to the seventh power? One would need a PhD in
irrational math to figure this out, certainly. So I thought, what
would Dr. Tannenbaum do? Research, of course! The results are below
in the proposed installer GUI for Rosetta Booster, but first:
For the archives, I also need to address some misconceptions that can
get in the way of providing a solution to a client. The perpetuation
of these myths insults and belittles the outstanding effort of
Apple's implementation.
Here are the Top Ten myths about Rosetta, followed by my proposed
GUI, specifically labeled to insure that large numbers validate the
fears of the OP, like FOX News might do.
#1 " Rosetta is an emulator. "
No. Rosetta is a translator. Please read Tannenbaum, Apple, et al.
#2 "It uses up a lot of memory when running even if no ppc apps are
running."
As it must, of course. There is no free launch. However, this is
either the "Princess and the Pea" or "Achilles Heel" depending on
your system load and capacity for juvenile literary references. Apple
states that it allocates a large buffer. Both Apple's "large" and the
OP "a lot" are far too vague as they are relative to different
context. This is a pitfall of subjective thought. In comparison to
the size of the PPC code, the buffer is gynormous, humungous even! In
comparison to available resources, however, one could use miniscule
without challenge. On the other hand, if your system is at 90% used
RAM, you need to throw another RAM on the barbie, mate! Such memory
starved systems are not properly configured for the task. Properly
configured systems handle the load, as Atlas and turtles do.
Perspective is required here, to avoid mistaking an elephant's tale
for a snake or a rope. Please read Dr. Suess, Aesop, SnaggleTooth,
Crocodile Dundee, et al.
#3 "It also consumes cpu cycles just by virtue of running."
Whatdaya want for nuthin? Rubber Biscuit? Please listen to Blues
Brothers, et al.
#4 " fact that it remains running once started"
Only until the translated PPC process exits. Please read Apple, Dr.
Pepper, Chicken Little, et al.
#5 "No user wants their computer slowed down every time they boot. "
- Except the user annoyed at waiting for the OP game to launch first
time.
- Except for the power user that understands how operating systems work.
- Except for those users that hardly ever need to boot.
- Except for those users with boot times scheduled by OSX for
anytime before they arrive.
- Except the user that has PAID for enough RAM to run a typical mix
of applications.
- Except for users mounting Firewire, USB or file shared volumes
during startup
- Except for users with email and browsers as log in items.
- Except for users that, while the system boots, is concerned about
coffee and the phone message blinking.
Please consider actual user surveys and working with real users and
focus groups for credible input.
#6 No numbers 6-10, waiting for the OP to pull them out of the
original source of 1 through 5.
Now for ROSETTA BOOSTER!!!!:
Note that the figures are actual measurements. Source project FREE
off list, just ask.
NOTE - the gui screen shot (my BIGGGG punch line) was too big
(really, really big 27KB) to include on list, however, it was fun.
Reply off-list for just the screenshot.
The results:
exit(0) used an 33MB buffer for a 9KB app. When using our TCP RDBMS
app (33KB), it needed 35MB. Humungous!
On a 4GB machine:
Boot time increased 2.533 seconds or +6.535% for a 39 second boot.
Wired memory loss is 0.028% for a loss of one million one hundred
seventy thousand bytes
Active memory loss is 0.807% for a loss of thirty three million
eighty thousand swappable bytes
Inactive memory loss is 0.029% for a loss of one million two hundred
ten thousand swappable bytes
Princess and the Pea or Achilles Heel, your choice.
We had a lot of debate about what to name Rosetta Booster. Some
wanted to call it Rosetta Master, until marketing realized we were
snickering whenever they said "Rosetta Master Beta". Reference Beavis
& Butthead, et al.
Enjoy!
Gary
"As old as Tannenbaum"
Not sent from my iPhone.
_______________________________________________
Cocoa-dev mailing list (email@hidden)
Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden