Fascinating thread guys, but can we close the book and move on now?
Thanks!
*********** REPLY SEPARATOR ***********
On 2/1/00 at 10:34 PM Randy Stafford wrote:
>Tim Endres wrote, quoting me:
>
>> > I develop on Windows. I have compared your jCVS 4.7 and wincvs.org's
>> WinCVS
>> > 1.0.6 and found jCVS to be vastly inferior, and I'm cross-posting to
>> > [EMAIL PROTECTED] and comp.software.config-mgmt. I'm recommending WinCVS
>> to
>> > everyone. I don't have a legal or ethical problem because both jCVS and
>> > WinCVS are distributed under the GNU Public License ("activities other
>> than
>> > copying, distribution and modification are not covered by this License;
>> they
>> > are outside its scope").
>>
>> Sounds fine. I would post a reply stating that people should try 5.1.
>> It competes quite well with WinCVS.
>>
> [Randy Stafford] Ah. So the playing field wasn't exactly level,
>because I compared the most recent version of a competing product with an
>older version of your product. You'd feel compelled to make that point in
>defense of your product's reputation and that of its developer(s). But by
>that time, some damage might already have been done, because the tilted
>comparison would have been made public, and could potentially influence the
>thinking of future technology selectors who might not have the opportunity
>or inclination to consider your rebuttal.
>
> This is the position that GemStone finds itself in. The Charles
>University EJB server comparison published last week used GemStone/J 3.0.2,
>which was outdated four months ago by GemStone/J 3.1 (itself outdated
>yesterday by GemStone/J 3.2). Let me quote to you from the opening
>paragraphs of the release notes of GemStone/J 3.1: "the principal focus for
>the 3.1 release of GemStone/J has been to optimize the product for
>performance, scalability, and reliability. We've made a large number of
>improvements in all of those areas... Performance has been improved across
>the board in GemStone/J 3.1... In general, we have improved internal
>designs, reduced unnecessary internal communications, optimized algorithms,
>and removed unnecessary code. We expect just about every kind of
>application to see noticeable performance improvements from this release,
>without any changes to application code." I have personally observed 3X
>improvements between 3.0.2 and 3.1 in the application I'm working on - out
>of the box. And 3.2 uses the HotSpot VM by default, which further improves
>matters. Among other failures, the Charles University comparison fails to
>account for the performance improvement effects of these releases.
>
>> Where is the problem, Sir?
>>
>> Only in your *intentional* spreading of misinformation. Which has nothing
>> to do with an honest, third party, details fully disclosed test.
>>
> [Randy Stafford] I want you to know that I have not posted any CVS
>client comparison conclusions on [EMAIL PROTECTED] or
>comp.software.config-mgmt, nor do I intend to. I was merely hypothesizing
>an analogous situation to illustrate the above point. I can see by the
>level of concern evident in your reaction that you care about your product's
>(and company's) reputation, and you care that it is fairly evaluated and
>publicly critiqued. So do we, in regards to our product. No one would
>expect anything less of either of us.
>
> Frankly I have never undertaken a serious evaluation/comparison of
>GUI CVS clients. I did try jCVS a little less than two years ago, but
>abandoned it for the command line. I didn't have a major problem with it
>per se; it just seemed that the GUI was a little clunky and I was more
>efficient on the command line - therefore jCVS was useless to me. And I am
>currently using WinCVS, because it adds value while still providing a
>command-line mode and is therefore useful.
>
> But suppose I did undertake a comparative evaluation of GUI CVS
>clients. Suppose I evaluated the candidate clients on a machine with too
>little memory to accomodate the Swing-induced footprint increase of jCVS 5.1
>over jCVS 4.7.5's AWT footprint (acknowledged in the 5.1 release notes), and
>suppose I concluded that 5.1 performed poorly as a result. Suppose some of
>my criteria were support for arcane CVS commands that are never really used
>in practice and might someday be phased out, and suppose jCVS got low marks
>in this area (I don't know that it would). Suppose I was over-emphasizing
>tight Windows integration and naively ignoring the broader set of usage
>scenarios afforded my be the portability of jCVS. Suppose I published my
>conclusions without first giving you an opportunity to point out these
>issues, and point out to me how jCVS 5.1 "competes quite well with WinCVS",
>as you put it. The result is that damage might be done to your product's
>and company's reputation. For better or worse, this is the kind of thing
>against which the license clauses to which you object are intended to
>protect. The Charles University EJB server comparison contains analogous
>shortcomings with respect to its treatment of GemStone/J, as pointed out the
>other day by Lougie Anderson, our VP of Engineering.
>
>> Frankly, I will place my lot with my code. Which has proven to be quite
>> a successful strategy to date.
>>
> [Randy Stafford] As will we. GemStone has been shipping object
>application servers since 1987, business has never been better, and our
>trajectory will continue to sustain and improve our position in the market.
>
>> > PPS. Your domain name implies that you want people to trust you, and
>> yet
>> > you're openly announcing that you'll litigate the validity of software
>> > licenses?
>>
>> Randy, please, I am feeling bad for you now. I hope that people would
>> trust
>> me not to lie down and take it from an overzealous ministry, er, I mean
>> vendor.
>> I guess it depends on your definition of trust. I trusted that report to
>> remain
>> available on the internet, and now it is gone.
>>
> [Randy Stafford] Don't bother feeling bad for me - it's wasted
>emotion. Obviously you're not the type to lie down, and we're not trying to
>make you or anyone else "take it". Nor are we an "overzealous ministry" -
>we're just defending the reputation of our product and company, as you did
>with your reply to my message. And we're just as proud of the Smalltalk
>tradition from which we come as you are of the Unix and open source
>tradition from which you come. It's obvious that you are contributing a
>great deal to the Java community, as we feel we are with some of our unique
>features, and that's a noble thing. We would all do well to remember that
>we are part of the same community - would we rather be programming DCOM?
>
> One last thing - you seem to be making an implicit assumption in
>this and other posts that we had something to do with removing that report
>from the Internet. As far as I know we did not. My best guess is that
>Charles University removed the report, perhaps to redress its shortcomings
>which should by now be obvious, or perhaps out of responsible legal concern
>over potentially violating not only GemStone's but also WebLogic's license
>agreement.
>
> Randy Stafford
> Senior Architect
> GemStone Systems, Inc.
>
>===========================================================================
>To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>of the message "signoff EJB-INTEREST". For general help, send email to
>[EMAIL PROTECTED] and include in the body of the message "help".
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
