Are the performance criteria outlined a few months ago by JTK, polished up
by Gerv and Jesus X and webbed by Hixie under any consideration for Mozilla
1.0?
(The idea being, I presume, that stable APIs are not enough, the reference
implementation must also work usably.)
http://www.damowmow.com/mozilla/mozilla-1.0.txt
Here's that page:
Proposed Mozilla 1.0 Release Criteria
=====================================
The final release criteria must be published well prior to release.
USABILITY
All bugs marked with the following markers in the status whiteboard
should be fixed: [Hixie-P0]
PERFORMANCE
The goals of our performance criteria are for a user who switches from
the at-that-point current version of IE to the Mozilla 1.0 build to
not experience a regression in performance that is greater than a
factor of two. (Ideally, of course, there would be no regression at
all. This is simply a goal for a 1.0 release.)
1. On any platform where IE also runs, the corresponding Mozilla build
will be no worse than twice as slow as the currently-released version
of IE at any task (for any task which takes measurable time [1]). This
includes things such as new window time, page loading time, showing
the bookmarks menu, find in page, and so on.
2. Memory usage will meet the same criteria as #1, i.e., Mozilla will
use no more than twice the amount of memory that IE requires to do the
same task. Again this applies to the UI as well as the HTML rendering.
3. Mozilla's mail/news reader will meet the same requirements as in #1
and #2, but with respect to Netscape Communicator 4.77.
4. Memory leaks will be no greater than twice the leakage of IE or NC
4.77, whichever is greater (or applicable, eg. in the case of the
mail/newsreader), for the equivalent operation(s).
5. True startup time of the application (from a cold start, i.e.,
without any "-turbo" cheating) will be no longer than two times the
combined startup times of Netscape Communicator 4.77 and the current
(at time of Mozilla release) version of IE. That is, t(M) <=
2*(t(NC)+t(IE)).
6. All these release criteria will be tested for compliance by
publically-available, published, and reproducible-by-anybody-
who-wants-to means.
7. "Official" performance test results will be published at the time
of release. "Units" will be in "IEs" or "NCs", as appropriate, so
potential downloaders can easily form an idea as to the performance to
expect.
Notes:
[1] The use of statistical methods to measure the time a task takes to
complete, such as loading 100 pages to determine the average page load
time, means that the time for such tasks is not immeasurable.
STANDARDS COMPLIANCE
1. All bugs marked with the following markers in the status whiteboard
should be fixed: [Hixie-P1] [Hixie-P2] [Hixie-P3] [Hixie-P4]
2. All CSS1 bugs and RFEs required to claim full support and
conformance should be fixed.
3. All HTML4 bugs and RFEs required to claim full support and
conformance should be fixed.
4. All DOM1 bugs and RFEs required to claim full support and
conformance should be fixed.
5. All CSS2, CSS3, XML, DOM2, DOM3, HTTP, etc, bugs that break a
feature in such a way as to prevent adoption should be fixed (removing
the feature altogether counts as a valid fix for these bugs).
MINIMUM REQUIREMENTS
The criteria described in this document only needs to be met on
machines meeting or exceeding the official minimum requirements for
the Mozilla 1.0 release.
These requirements vary by platform.
On Windows, they are: 266MHz Pentium with 64MB RAM.
Requirements for Mac and Linux machines have yet to be determined.
CONTRIBUTORS
Thanks to: JTK, Jesus_X, Gerv
--
http://thingy.apana.org.au/~fun/ http://www.rocknerd.org/
"One of my remaining ambitions is to receive a blowjob while remotely
configuring someone's router." (Lionel Lauer)
