FAQ: Netscape Communications and our products
er...@rdr.com
>I have two more questions to add to your FAQ list:
>Q. How does NCC propose to make amends to the public for its
> broken promise to provide a "free for personal use" version of
> Netscape Navigator?
>Q. What measures has NCC undertaken that will reassure the public
> that NCC will behave with a higher degree of ethics and integrity
> in the future than it had behaved in regard to this "free
> for personal use" broken promise.
> I'll strongly second this and encourage others who feel this way to do the
same.
> Marc A's FAQ post is just one more example of Netscapes lack of respect for
> the Net. They refuse to awnser any questions that place them in bad light,
and
> when they do get around to posting something, they ignore the questions that
> have been raised.
> You do have a fine product (ignoring the code extensions) and there is no
question
> that it is worht what you are asking for it. The questions involve the
integrety of
> the company and they need to be addressed.
>
> Bill Spurlock
>
> The one thing that Netscape has proven is that it has only as much
> integrity as the marketplace imposes on it. Obviously, the protests
> haven't hurt sales enough for NCC to respond.
> If they do, it will be to protect their bottom line - not their
> "integrity"
>
I have just recently started reading this newsgroup and have no idea of The
problems discuused here in this post. I have been using Netscape since the
093 beta and have not seen anything on Paying for it. Has there been some
press release from NCC... I do no that they have never responded to any of my
numerous e-mails over the last 2 months...
Eric Ruck
Joe Zychik
> problems discuused here in this post. I have been using Netscape since the
> 093 beta and have not seen anything on Paying for it. Has there been some
> press release from NCC... I do no that they have never responded to any of my
> numerous e-mails over the last 2 months...
>
> Eric Ruck
>
>
>>>>
1.0 free for personal use. Unless you are an employee of a non-profit organization
(which includes white-supremecist organizations that have nonprofit status) or you
are an educator or student you are REQUIRED to pay $39 to use Netscape.
Your other option is to take full advantage of the license and evaluate 1.0
for as long as you wish! This option, IMO, was placed in their by their lawyers
so that you could still use it for "free" as long as long as you were willing
to sink to their devious level.
jz
a nobody.
___________________________________
"Freedom Sux"
Josef Stalin
"You don't have to think. I'll do it for you."
Famous Radio and TV Supreme Commander Dittohead
Dan Empfield
I guess I have another question. I registered my product and gave you my
Visa# and you haven't taken my $39 yet. Wondering when you will.
You can have the $39 with no reservation, and I agree with even those who
have so vehemently taken umbrage with you, that regardless of your level
of integrity $39 is a great price for the product.
Here is my un-FAQ, not for NCC, but for the rest of you:
Q: If NCC was to do as many of you propose,to continue to let those of you
who helped test the original version of this product use it N/C, how
exactly would NCC ever realize a profit from this project?
A: ?
(It would seem to me, slow witted and uninitiated as I am, that the
product would be reduced to sharehware ("If you are using this product
after xx/xx/94 please send $39"). Maybe I'm missing something. Maybe you
can make a fortune off of selling server software. I don't know.
Dan Empfield
Harry J. Saal
>
>Some FAQs -- Netscape Communications Corp.
>Version 0.3; last modified 1/10/94, ma...@mcom.com.
>
>
>Q. What's a nonprofit organization? Does a government agency count
> as a nonprofit organization?
>
>A. A nonprofit organization in the United States is an entity to whom
> the Internal Revenue Service has granted 501(c)(3) status.
> We don't currently have a straightforward answer for non-US organizations.
> Government agencies are not considered nonprofit organizations.
>
This is an incorrect answer, IMHO. Smart Valley, Inc. is a nonprofit
organization. The IRS has granted us 501(c)(6) status, e.g. there
are multiple forms of non-profit. 501(c)(3) are typically charities
although there are numerous organizations such as CommerceNet
that applied for 501(c)(3) status and received it.
So is the free licence to non-profits as the IRS defines them or
as this FAQ attempts to clarify? The language in the either
the licence or this FAQ ought to be changed.
============================================================
Dr. Harry J. Saal, President & CEO, Smart Valley, Inc.
http://www.svi.org/ gopher://gopher.svi.org
E-mail: HS...@svi.org Voice: 415-843-2102 FAX: 415-843-2222
============================================================
Lloyd Zusman
> [snip]
> Here is my un-FAQ, not for NCC, but for the rest of you:
>
> Q: If NCC was to do as many of you propose,to continue to let those of you
> who helped test the original version of this product use it N/C, how
> exactly would NCC ever realize a profit from this project?
>
> A: ?
>
> (It would seem to me, slow witted and uninitiated as I am, that the
> product would be reduced to sharehware ("If you are using this product
> after xx/xx/94 please send $39"). Maybe I'm missing something. Maybe you
> can make a fortune off of selling server software. I don't know.
In my opinion, here's what NCC could have done, and still could do:
(1) Publicly apologize for the broken promise about version 1.0
being "free for personal use".
(2) Explain to the net community how they made an error in judgment
in thinking that their company could still be profitable if
version 1.0 was free for personal use, and how they had to
do *something* to bring in some money once version 1.0 was ready
to ship (assuming that this is indeed what happened).
(3) Publicly acknowledge that NCC has profited by the good will
and support that they got from people on the net who
promoted and helped test the product under the assumption
that version 1.0 would be "free for personal use".
(4) Offer a free copy of version 1.0 to every person on the net
who can demonstrate that he or she provided tangible help
in beta testing or promoting the product prior to the release
of version 1.0. This should be easy to verify by checking
comp.infosystems.www.* news archives and by looking at the
email that was received and the logs that were kept by NCC's
tech support people. I estimate that this would apply to no
more than a dozen or so people on the net.
(5) Offer some other gift to these people, as well ... perhaps
free advertising space on the NCC Web pages, or something.
(6) Swear publicly and humbly that they have learned their lesson
and will not make rash promises again. Make it clear that
they want to work with the net community not only to continue
to make Netscape Navigator a still better product, but also to
help make their company a more responsive organization.
This would probably cost NCC no more than $1000 or so, and it would go
a long way towards making amends for the broken promise. In my opinion,
this would be a much more ethical and honest way of dealing with the
situation NCC found itself in as it was ready to release version 1.0.
Note that this is not the same as offering their product as shareware
or freeware.
--
Lloyd Zusman 01234567 <-- The world famous Indent-o-Meter.
l...@panix.com ^ I indent thee.
To get my PGP public key automatically mailed to you, please
send me email with the following string as the subject or on a
line by itself in the message (leave off the quotation marks):
"mail-request public-key"
Joe Zychik
They could offer 1.0 and two free upgrades.
They could win a lot of sentiment and support by asking for suggestions on
how they could rectify the situation. Then Lloyd and I can argue with
each other over what's the best way to rectify things. And he can tell me
I take cheap shots and I can ignore him :-)
Nicolas Pioch
[comp.infosystems.www.misc]
| (1) Publicly apologize
| (2) Explain to the net community how they made an error in judgment
| (4) Offer
| (5) Offer some other gift
har har har!
(7) Send a box of chocolate cookies to all comp.infosystems.www.*
whiners too maybe ? :)
They've been giving a great product free for internet use the last
two months already! I'm still waiting to see that much from the
competitors.
-- Nicolas
Sima Desai
>If you are using 1.0 read the license. NCC has gone back on their promise to make
>1.0 free for personal use. Unless you are an employee of a non-profit organization
>(which includes white-supremecist organizations that have nonprofit status) or you
>are an educator or student you are REQUIRED to pay $39 to use Netscape.
Here's the README.TXT file from .94.... Or at least the beginning of it:
==============================================================================
Netscape 0.94 beta (Windows)
==============================================================================
Netscape 0.94 (public beta version) is free for your personal use,
subject to the terms detailed in the license agreement accompanying it.
Subject to the timing and results of this beta cycle, Netscape Communications
will release Netscape 1.0, also available free for personal use via the
^^^^^^^^^^^^^^^^^^^^^
Internet. It will be subject to license terms; please review them when and
if you obtain Netscape 1.0.
A commercial version of Netscape 1.0, including technical support
from Netscape Communications, will be available upon completion of the
beta cycle.
******
So anyone who said (and at least two people did) that Netscape never said
1.0 would be free: you're a ninny.
Sima Desai
>Q: If NCC was to do as many of you propose,to continue to let those of you
>who helped test the original version of this product use it N/C, how
>exactly would NCC ever realize a profit from this project?
A1: 2.0 with new whiz bang feateres.
A2: Selling thousand's of servers for $5000 each outta do it.
A3: Who gives a rat's ass, paying for their promises is their problem.
>after xx/xx/94 please send $39"). Maybe I'm missing something. Maybe you
>can make a fortune off of selling server software. I don't know.
My perception is that Netscape purposely released a product that was a
total BITCH on servers (multiple connections, one for the html document,
and another for each and every inline gif. Current servers fork off a
separate process for each conncetion) with the intent of making a whiz-
bang server that doesn't fork off a separate process. Netscape servers
would then be NEEDED because other servers would be comparably bogged
down by greedy, resource-grabbing Netscape clients. Netscape has a
great server, it's fast as hell becasue it doesnt fork off a process
for each connection, which entails huge overhead. It's really a great
scam, except that Netscape is millions in the hole and desperate for
cash right about now. So they backed out on free copies for personal
use.
I think Netscape has a few not-perfectly-clean schemes going on,
but I don't think they were ever thinking "Let's tell they can have it
for free to get free beta-testing." Maybe I'm wrong.
Sima Desai
>(7) Send a box of chocolate cookies to all comp.infosystems.www.*
>whiners too maybe ? :)
You must have been lied to often, by your parents or something,
becasue you take it way to easy. Me, I get pissed when people lie
to me.
>They've been giving a great product free for internet use the last
>two months already! I'm still waiting to see that much from the
>competitors.
NCSA Mosaic has always been free.
hawkins
thom...@robin.tezcat.com says...
>Do you or anyone have an actual verbatim copy (not your interpretation of
what
>they said) of the "promise" to provide a version of Netscape free for
personal
>use in perpetuity? If so, please post. Personally, I'll believe it when
I
>see it.
>/bob/
Believe me .... the Netscape people don't deny they said this ...
They just changed the rules .... if you read there FAQ you'll see
they attempt to explain the situation.
And STILL NO WORD ON BUG FIXES BEING COVERED OR NOT BY THE $39 fee!!!!
Netscape .... pls phone home!!
Bob Thompson
>From: l...@panix.com (Lloyd Zusman)
>Subject: Re: FAQ: Netscape Communications and our products
>Date: 10 Jan 1995 07:30:12 -0500
>[ This is being posted to the comp.infosystems.www.misc newsgroup,
> and CC's are being sent to sa...@mcom.com and ma...@mcom.com ]
>In article <marca-10019...@boulanger.mcom.com>, ma...@mcom.com (Marc
>Andreessen) writes:
>> Some FAQs -- Netscape Communications Corp.
>> Version 0.3; last modified 1/10/94, ma...@mcom.com.
>I have two more questions to add to your FAQ list:
>Q. How does NCC propose to make amends to the public for its
> broken promise to provide a "free for personal use" version of
> Netscape Navigator?
Do you or anyone have an actual verbatim copy (not your interpretation of what
Jon Mittelhauser
> True, but how many fork()ed httpds would you have at any given time?
> When Netscape opens multiple connections, the httpd fork()s off a
> number of processes which reply concurrently. With Mosaic (or for that
> matter, any other browser), the httpd fork()s off processes
> sequentially because the browser waits until one request has been
> answered before sending the next.
Yes. But you are only thinking of ONE user. You need to extend
your graph to 3 dimensions (multiple users over time). With
Netscape, the httpd forks 4 connections but also finsihes them
first. I've drawn this graph before but I will again:
NCSA:
AAAABBBBCCCCDDDD
AAAABBBBCCCCDDDD
AAAABBBBCCCCDDDD
AAAABBBBCCCCDDDD
NCOM:
AAAABBBBCCCCDDDD
AAAABBBBCCCCDDDD
AAAABBBBCCCCDDDD
AAAABBBBCCCCDDDD
This is an oversimplification (but you can extrapolate). We have
4 users who each start a 4 document connection at 1 time slice
intervals. If you extend both graphs over time, you will see
that both servers sit at an average of 4 connections/time slice.
If you are dealing with something as complex as server load you
cannot estimate it by looking at a single user! Even looking at
these 3 dimenstions is not really enough to do it justice...
experience is the best teacher...
-Jon
Nicolas Pioch
[comp.infosystems.www.misc]
| >Having multiple connections does *not* change the total number of
| >connections you know... actually this total number DECREASES because
| >of the client-side caching!
|
| Caching decreases the number of server accesses. Having multiple
| concurrent processes for each client is a stress on the server.
| What Netscape should do is read all the text in first, displaying
| it as it comes in, THEN start d/ling images. Not all at the same time.
That's what Netscape does provided you use the width= and height=
attributes (I know other browsers using them too).
If you don't use them, well... to be able to layout the pages, you
need to download the first bytes of any image!
| >This was for the provider issue... for end-users, having multiple
| >connections is clearly a winning strategy since it allows data display
| >on the fly!
|
| Sure, clients dont care that they've forked off 5 process at one time on
| the server. They dont pay for the CPU cycles.
You're biaising the discussion, since you're trying to make people
believe that Netscape *increases* the CPU cycles, which is NOT the case.
Re-read my previous post carefully, and you'll see that
- netscape does NOT increase the total # of CPU cycles or # of fork()s.
Just the *timely repartition* gets changed (and statistically it doesn't
change much as soon as you're having little busy server)
- actually netscape (as chimera and other caching systems)
DOES SAVE cpu cycles, because of the disk caching.
Once again: Netscape, Chimera, CERN httpd (in proxy/caching mode)
all save a tremendous bandwidth, both to providers and end-users.
Hope that's clearer now,
-- Nicolas
Bill Spurlock
>If you are dealing with something as complex as server load you
>cannot estimate it by looking at a single user! Even looking at
>these 3 dimenstions is not really enough to do it justice...
>experience is the best teacher...
Uh Jon, this is all well and nice, but would you care to explain
why you can deal with this and not reply to my question?
k claffy
>First, the Netscape client's method of parallel loading was designed
>to improve the user's perception of the speed of loading WWW
>documents. The method _does not_ result in a higher load for
>servers. They are still transferring the identical amount of
>information. Yes, it does potentially increase the maximum load that
>a server will carry at any one time. However because of the nature
>of the Web, the load over time becomes the critical measure and
>Netscape doesn't affect that. A number of the major sites on the
>Web were initially concerned when this issue was raised but
>experience has shown that we were correct and they have confirmed
>that Netscape's methods were not "a BITCH" on servers. You might
>walk next door and talk to some of the people at CCSO (or even
>NCSA) and ask them yourself if you doubt me.
Jon,
btw i was wondering on what you based
this default number of 4 allowable
connections at once? did you do
some performance testing to
determine that was optimal,
or was it just a hunch?
has anyone else out there tried
other values with noticeably different
results?
[the online doc indicates that one can
only change the default on a macintosh,
is this true? strange -- ]
>In fact, Netscape is _extremely_ resource friendly to severs.
>We are seeing over 30% of the hits on our server resulting in
>304 (use cached copy) messages. Friday (1/13) alone, this
>resulted in almost 2.5 GIGAbytes of information which didn't
>have to be transfered. This is good on both servers and the
>network itself.
uh, where exactly is that stuff
being cached?
tnx
k
Sima Desai
>This is an oversimplification (but you can extrapolate). We have
Sure is.
>4 users who each start a 4 document connection at 1 time slice
>intervals. If you extend both graphs over time, you will see
>that both servers sit at an average of 4 connections/time slice.
Those four connections an NCOM user has will be open longer than NCSA
Mosaic's four connections. Having four connections doesnt make the
user's 'pipe' four times as thick. Having four connections through
one pipe means each four connections are open 4 times as long. This too
is an oversimplification, as all four connectiosn wont be opened
simultanesouly, but rather when the new document is run into in the
main document, and not all the documents are the same lenght, but...
Your time/#processes graph should look more like this:
# Connections is Y; Time is X:
NCOM
____
AAAA
BBBB
CCCC
DDDD
NCSA
____
ABCD
This too can be extrapolated over as many users as you want. The same
amount of 'stuff' is coming out over the net, yes, but the load on the
server is higher, because it has to maintain a higher number of simultaneous
connections.
Leon D. Shaner
> >4 users who each start a 4 document connection at 1 time slice
> >intervals. If you extend both graphs over time, you will see
> >that both servers sit at an average of 4 connections/time slice.
>
> Those four connections an NCOM user has will be open longer than NCSA
> Mosaic's four connections. Having four connections doesnt make the
> user's 'pipe' four times as thick. Having four connections through
> one pipe means each four connections are open 4 times as long. This too
> is an oversimplification, as all four connectiosn wont be opened
> simultanesouly, but rather when the new document is run into in the
> main document, and not all the documents are the same lenght, but...
Your assumption is that each of those four connections takes 25% of the
bandwidth. Not necessarily the case.
Provided that all four connections combined take less than the total
bandwidth (very likely) then in fact the combined time for the parallel
connections is no longer than the combined time for the sequential
connections.
The only time the connections might hang around longer when done in
parallel is when the client machine is dog slow and can't really process
things as efficiently as the bandwidth should allow.
--
---
Leon D. Shaner * http://www.cts.mtu.edu/~leon * (906)487-2112
Computing Technology Services * Michigan Technological University
Barry Margolin
>Yes. But you are only thinking of ONE user. You need to extend
>your graph to 3 dimensions (multiple users over time). With
>Netscape, the httpd forks 4 connections but also finsihes them
>first. I've drawn this graph before but I will again:
>
>NCSA:
>
> AAAABBBBCCCCDDDD
> AAAABBBBCCCCDDDD
> AAAABBBBCCCCDDDD
> AAAABBBBCCCCDDDD
>
>NCOM:
>
> AAAABBBBCCCCDDDD
> AAAABBBBCCCCDDDD
> AAAABBBBCCCCDDDD
> AAAABBBBCCCCDDDD
>
>This is an oversimplification (but you can extrapolate). We have
>4 users who each start a 4 document connection at 1 time slice
>intervals. If you extend both graphs over time, you will see
>that both servers sit at an average of 4 connections/time slice.
Your oversimplification neglects the fact that transactions go faster when
there are fewer simultaneous transactions. A more accurate picture of the
NCSA scenario would be:
A---B---C---D
A---B---C---D
A---B---C---D
A---B---C---D
I admit that this is also an oversimplification: it assumes that the
performance degradation will be linear. It also gets more complicated if
the users each start their document connections at <1 time slice
intervals. For instance, if they start at .25 time slice intervals, it
looks more like:
ABBCCCDDDD
AABBBCCCCDDD
AAABBBBCCCDD
AAAABBBCCD
If steady state is reached each document will take four time slices
to be downloaded, and the pattern will look like your NCSA diagram.
Of course, this comparison is really only interesting for servers that are
being accessed *very* frequently (i.e. continuously most of the time), like
the default home pages used by the popular browsers and the "Starting
Points for Internet Exploration". Most servers aren't going to get so much
overlap in accesses from different users that they'll get into that steady
state I described.
--
Barry Margolin
BBN Internet Services Corp.
bar...@near.net
Jochen Bern
>Here is my un-FAQ, not for NCC, but for the rest of you:
>Q: If NCC was to do as many of you propose,to continue to let those of you
>who helped test the original version of this product use it N/C, how
>exactly would NCC ever realize a profit from this project?
The initial Announcements from NCC basically said "we'll make our
Profit from selling the SERVER Software and doing Servicing".
Another Possibility that's always been discussed (as witnessed
by the Concern about the "jettison X Days after new Version"
in the Beta Licenses) is to give the first Version(s) for free
and start selling the more elaborated later ones.
Regards,
J. Bern
--
/ \ Mail: be...@Uni-Trier.DE (7Bit); be...@TI.Uni-Trier.DE (OW Mails OK); /\
/ J. \ <A HREF="http://www.informatik.uni-trier.de/~bern/">My Homepage</A> / \
\Bern/ FINGER, NEWS ETC. AT THIS SITE ARE BROKEN, PLEASE USE MAIL (+ WWW)! \ /
\ / P.O. Box 1203, 54202 Trier, Germany, Europe -- Ham Call Sign: DD0KZ \/
Jochen Bern
> Netscape Navigator
> and our Netsite server line is aggressively standards-compliant
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
... after having defined a new HTML Standard itself to comply to ...
Joe Buck
>When Netscape opens multiple connections, the httpd fork()s off a
>number of processes which reply concurrently. With Mosaic (or for that
>matter, any other browser), the httpd fork()s off processes
>sequentially because the browser waits until one request has been
>answered before sending the next.
If you only have one user at any given time, then yes, your distribution
of forked processes at any given time would be burstier with Netscape,
and you'd see a peak of four active connections rather than one.
But if your site is busy enough so on the average, several different
clients are actively accessing your site at any given time, that
burstiness is going to get smoothed out. So Netscape's behavior makes
little difference for highly loaded sites. For lightly used sites,
it certainly increases the peak load.
--
-- Joe Buck <jb...@synopsys.com> (not speaking for Synopsys, Inc)
Phone: +1 415 694 1729
Jon Mittelhauser
Because you are not asking an engineering question.
I'm trying to get an "official" response. I am not qualified to
give you one. All I can say is that there was no malicious
intent in the change of policy. External factors dictated the
change and if we can get an official explanation, I will post
it.
-Jon
Jon Mittelhauser
> Jon,
> btw i was wondering on what you based
> this default number of 4 allowable
> connections at once? did you do
> some performance testing to
> determine that was optimal,
Yes. It, however, does vary with your particular configuration.
> >In fact, Netscape is _extremely_ resource friendly to severs.
> >We are seeing over 30% of the hits on our server resulting in
> >304 (use cached copy) messages. Friday (1/13) alone, this
> >resulted in almost 2.5 GIGAbytes of information which didn't
> >have to be transfered. This is good on both servers and the
> >network itself.
>
> uh, where exactly is that stuff
> being cached?
To the users local disk.
-Jon
Tim Lahey
wrote:
> In <marca-10019...@boulanger.mcom.com> ma...@mcom.com (Marc Andreessen) writes:
> > Netscape Navigator
> > and our Netsite server line is aggressively standards-compliant
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> ... after having defined a new HTML Standard itself to comply to ...
>
> Regards,
> J. Bern
It is typical practice in areas where the standards are not well developed,
or
even in areas where standards are well developed to first implement the
idea
then get it standardized. Seeing as how I am in a standards department as
well
as having worked in development and planning I have seen implementations
drive
standards many times. So instead of criticizing Netscape Comm. you should
thank them for adding new features to the standards.
Regards,
Tim Lahey
---
Note: I speak for myself not Bell-Northern Research or anyone else for
that matter.
Bill Spurlock
>I'm trying to get an "official" response. I am not qualified to
>give you one. All I can say is that there was no malicious
>intent in the change of policy. External factors dictated the
>change and if we can get an official explanation, I will post
>it.
You'll understand if I don't hold my breath waiting for this official
explanation.
At the very least, thank you for your reply.It's nice to know that
someone ant NCOM will say something regarding this mess.
(And he did not ask me who I am...:) )
Bill
Joe Zychik
ask you something. Are you aware that NCC promised that 1.0 would
be free for personal use? I'm not trying to be facetious. I'm trying to
get a dialogue going. Hopefully we can at least agree on some facts
and then take it from there.
sincerely,
jz
Tom Neff
Sima Desai <sd9...@ehibm5.cen.uiuc.edu> wrote:
>This too can be extrapolated over as many users as you want. The same
>amount of 'stuff' is coming out over the net, yes, but the load on the
>server is higher, because it has to maintain a higher number of simultaneous
>connections.
Actually, on a saturated server, you CAN'T have a higher number of
simultaneous connections; all you can do is monopolize a bigger fraction
of the available connections for a single client. This is why
Netscape's multi-connection behavior is rude, and their attempt to sell
it as "better for the server" is disingenuous. Mozilla is not the only
offender, by the way; my own main squeeze (WebExplorer) can go five wide
on a big, graph-rich page -- 1 for the doc and 4 for images. I don't
know what Netscape's upper limit is, if it even has one.
--
Tom Neff [email] tn...@panix.com [web] http://www.panix.com/tneff/
Jon Mittelhauser
>
> In article <3fgmg9$i...@vixen.cso.uiuc.edu>,
> Sima Desai <sd9...@ehibm5.cen.uiuc.edu> wrote:
> >This too can be extrapolated over as many users as you want. The same
> >amount of 'stuff' is coming out over the net, yes, but the load on the
> >server is higher, because it has to maintain a higher number of simultaneous
> >connections.
>
> Actually, on a saturated server, you CAN'T have a higher number of
> simultaneous connections; all you can do is monopolize a bigger fraction
> of the available connections for a single client. This is why
> Netscape's multi-connection behavior is rude, and their attempt to sell
> it as "better for the server" is disingenuous.
As has be pointed out in earlier replies, the time factor levels
out the peaks for a heavy loaded server. This has been shown in
both theoretical but more importantly _real life actual_ experience.
For a lightly loaded server, the peaks would be higher but if the
server isn't loaded, who cares?
I never represented that the multi-connection behavior was better
for the server. I represented that the client as a whole was better
for HTTPd servers. That was based on the lower bandwidth caused
by the disk cache. The multi-connection issue is orthoganal.
-Jon
Ed Kubaitis
>
> In article <3fhigk$d...@flop.mcom.com>, Jon Mittelhauser <jo...@mcom.com> wrote:
> >sha...@mindspring.com (Bill Spurlock) wrote:
> >I'm trying to get an "official" response. I am not qualified to
> >give you one. All I can say is that there was no malicious
> >intent in the change of policy. External factors dictated the
> >change and if we can get an official explanation, I will post
> >it.
>
> a bullshit attitude that it takes _weeks_ to get a response to a
> fundamental ethical question. I consider it a complete insult that it
> could take this long to get an answer.
>
> Paul Prescod
On the other hand, *I'm* starting to get really irritated
by Netscape's lack of killfiles.
--------------------------
Ed Kubaitis (e...@uiuc.edu)
Computing & Communications Services Office - University of Illinois, Urbana
Paul Prescod
>sha...@mindspring.com (Bill Spurlock) wrote:
>give you one. All I can say is that there was no malicious
>intent in the change of policy. External factors dictated the
>change and if we can get an official explanation, I will post
>it.
It boggles my mind that your little company already has developed such
Matt Ranney
[...]
: First, the Netscape client's method of parallel loading was designed
: to improve the user's perception of the speed of loading WWW
: documents.
And that it does. Quite nicely, in fact.
: The method _does not_ result in a higher load for servers.
[...]
While speed of multiple connection can be argued both ways, if your
server machine runs out of memory because of the httpd processes that
are sucking up .5M apiece or so, things are slower. If 20 people all
suddenly decide they want to load up one of your pages that has at
least 4 images on it, you now have 80 connections open, and 80 httpds
using up half a meg == 40M of RAM required. If you have a 32M machine
as your server, this can really be a problem, resulting in a MUCH
higher load for servers. If your 32M machine only had to deal with 20
connections at a time, only 10M of RAM would be needed, and there
would be no problem.
--
Matt Ranney - m...@nec.com
"You know, I don't think theres a man, woman, or child alive today
who doesn't enjoy a lovely beverage." -DL
Nicolas Pioch
[comp.infosystems.www.misc]
| are sucking up .5M apiece or so, things are slower. If 20 people all
| suddenly decide they want to load up one of your pages that has at
| least 4 images on it, you now have 80 connections open, and 80 httpds
| using up half a meg == 40M of RAM required. If you have a 32M machine
| as your server, this can really be a problem, resulting in a MUCH
| higher load for servers. If your 32M machine only had to deal with 20
| connections at a time, only 10M of RAM would be needed, and there
| would be no problem.
If everybody on earth was to walk one meter east at the exact same time,
the spinning/rotation of the planet would be changed, probably
extincting millions of species and causing unmatched disasters.
Fortunately, statistics teach us that not everybody's doing the same
thing at the same time, and so you needn't worry when walking east :)
-- Nicolas
Nicolas Pioch
[comp.infosystems.www.misc]
| In <marca-10019...@boulanger.mcom.com> ma...@mcom.com (Marc Andreessen) writes:
| > Netscape Navigator
| > and our Netsite server line is aggressively standards-compliant
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
| ... after having defined a new HTML Standard itself to comply to ...
That's how things move forward in the real world.
But then, if you've waited until ANSI C to learn C, are waiting for
ANSI C++, and renounce to TCP/IP, FTP... waiting for ISO/OSI crap
and FTAM... good luck :)
-- N.
JO...@olis.lib.ox.ac.uk
sd9...@ehibm5.cen.uiuc.edu (Sima Desai) writes:
[Snip]
> My perception is that Netscape purposely released a product that was a
>total BITCH on servers (multiple connections, one for the html document,
If your worried about server load then I think some student at MIT was
working on a threaded version of NCSA HTTPD.
[Snip]
Nicolas Pioch
[comp.infosystems.www.misc]
Dunno about MIT, but the multi-threaded web server is "MDMA" I think.
(Multimedia Daemon for Multiple Access or whatever)
I guess there's some litterature/code about that on sunsite.unc.edu.
Note that Netsite/MDMA don't have much things in common:
MDMA requires Solaris, and doesn't have any of the
other Netsite facilities as far as I know (configuration through
administrative forms, security in the Commerce server required for
online transactions, gif/jpg dynamic adaptability and all other fancy
stuff, etc).
But it's cool that someone in the academic domain is trying to
catch up/innovate the performance point of view, quite critical
and only solved by Netsite (and possibly MDMA) right now.
-- Nicolas
Hal Beumer
> Paul Prescod
Dear Jon and Marc
Thank you very much for continuing to post info to the net about Netscape and
for providing the product to us in the first place. As a professional
software engineer who works for a living in the real world, I am acutely
aware of the cost of developing and maintaining software. There has long
been an attitude prevalent in what I hope is a small subset of readers of this
newsgroup that they should get everything for free and that your time is worth
nothing. Flaming these people is a complete waste of time. If they were
susceptable to reason, they would not be making these kinds of posts. The tone
of many of the posts I have read about Netscape have been bitchy and whiny. I
for one have been using it since it came out and mosaic before it and think it
is an excellent product. I am sure most of the people on the net feel likewise.
Having said all these wonderful things, the blinking N is not getting any
prettier :-)
Hal Beumer
Bill Spurlock
>Dear Jon and Marc
Hal, having read the above all I can say is you don't have a clue about what we are talking about here.
Paul Prescod
Check my address. I am at an educational institution. I _can_ get Netscape
for free. The fundamental problem is that people in the computer software
industry _lie_ all of the time. Blatantly. To our faces. Marc Anderssen
is an obvious example of someone who had done so recently.
Is this an acceptable state of affairs to you? Do you like being lied to?
Do you like taking every vendor's "word" as a lie? As a software professional
you should be ashamed of those who give us a reputation for flashy demos
and empty promises. I know I am.
Paul Prescod
Steven Grimm
>Is this an acceptable state of affairs to you? Do you like being lied to?
>Do you like taking every vendor's "word" as a lie?
There's a difference between a "lie" and a "mistake," though from reading
this thread you'd think nobody had heard of the latter. Netscape thought
they had a pricing structure that would work for them. They looked at it
more closely and found they were wrong.
Why must everyone assume there was some sort of premeditation or malice
involved? Do you really believe that Marc A. and company had a meeting
one day and said, "Okay, we're going to promise it to everyone for free,
but after we've used everyone for beta-testing, we'll spring THIS on them?"
That's just plain paranoid. There's no evidence that they didn't mean
what they said when they made the initial announcement.
Not every untrue statement is a lie. If it were, every company in the
industry would be a liar -- who ever meets their announced ship dates,
for instance?
-Steve
Lloyd Zusman
> In <D2vnK...@undergrad.math.uwaterloo.ca> papr...@undergrad.math.uwaterloo.ca (Paul Prescod) writes:
>> Is this an acceptable state of affairs to you? Do you like being lied to?
>> Do you like taking every vendor's "word" as a lie?
> There's a difference between a "lie" and a "mistake," though from reading
> this thread you'd think nobody had heard of the latter. Netscape thought
> they had a pricing structure that would work for them. They looked at it
> more closely and found they were wrong.
I, for one, do not refer to this as a "lie". I call it a "broken
promise".
> Why must everyone assume there was some sort of premeditation or malice
> involved? Do you really believe that Marc A. and company had a meeting
> one day and said, "Okay, we're going to promise it to everyone for free,
> but after we've used everyone for beta-testing, we'll spring THIS on them?"
> That's just plain paranoid. There's no evidence that they didn't mean
> what they said when they made the initial announcement.
I agree that NCC probably was not trying to be malicious or to
deliberately take advantage of other people's good will. I think that
the most likely explanation for their broken priomise is that they did
not think their original promise through very carefully when they
first made it, especially insofar as it would affect the company's
"bottom line." Once it was nearing release time for version 1.0, they
probably looked at the financial implications of making the product
"free for personal use" and said, "OOPS!"
I have publicly stated this here already.
> Not every untrue statement is a lie. If it were, every company in the
> industry would be a liar -- who ever meets their announced ship dates,
> for instance?
I agree that all untruths are not necessarily deliberate lies, and I
believe that NCC is not likely to have planned a deliberate lie.
However, there has been no offer on NCC's part to make amends for this
"broken promise", "error in judgment", or whatever you prefer to call
it, and NCC has been publicly quite silent about this issue when
confronted here on the net. NCC has definitely profited from the good
will and the technical support of people here on the net during the
beta release process of Netscape, and some of these people, by their
own admission, gave more support to NCC than they otherwise would have
given if they knew that Netscape would not be "free for personal use"
after all. NCC originally stated that it plans to make most of its
profits from sales of its Web server software while releasing version
1.0 of Netscape as a "free for personal use" product.
I, for one, did not give any support to NCC during the Netscape beta
process, and so I do not feel personally betrayed, and I do not want
any kind of recompensation or "freebie". If I ever use Netscape, I
will pay for it. I am not using it now, however, and I am encouraging
my clients and associates to not use this software either, because of
the manner in which NCC has handled this broken promise issue so far.
I have previously posted some suggestions as to how NCC could make
amends, and I'm reposting them below. This is one of probably several
ways that NCC could win back some support and good will here on the
net:
(1) Publicly apologize for the broken promise about version 1.0
being "free for personal use".
(2) Explain to the net community how they made an error in judgment
in thinking that their company could still be profitable if
version 1.0 was free for personal use, and how they had to
do *something* to bring in some money once version 1.0 was ready
to ship (assuming that this is indeed what happened).
(3) Publicly acknowledge that NCC has profited by the good will
and support that they got from people on the net who
promoted and helped test the product under the assumption
that version 1.0 would be "free for personal use".
(4) Offer a free copy of version 1.0 to every person on the net
who can demonstrate that he or she provided tangible help
in beta testing or promoting the product prior to the release
of version 1.0. This should be easy to verify by checking
comp.infosystems.www.* news archives and by looking at the
email that was received and the logs that were kept by NCC's
tech support people. I estimate that this would apply to no
more than a dozen or so people on the net.
(5) Offer some other gift to these people, as well ... perhaps
free advertising space on the NCC Web pages, or something.
(6) Swear publicly and humbly that they have learned their lesson
and will not make rash promises again. Make it clear that
they want to work with the net community not only to continue
to make Netscape Navigator a still better product, but also to
help make their company a more responsive and ethical organization.
This would probably cost NCC no more than $1000 or so, and it would go
a long way towards making amends for the broken promise.
--
Lloyd Zusman 01234567 <-- The world famous Indent-o-Meter.
l...@panix.com ^ I indent thee.
To get my PGP public key automatically mailed to you, please
send me email with the following string as the subject or on a
line by itself in the message (leave off the quotation marks):
"mail-request public-key"
Ed Kubaitis
>
> Check my address. I am at an educational institution. I _can_ get Netscape
> for free. The fundamental problem is that people in the computer software
> industry _lie_ all of the time. Blatantly. To our faces. Marc Anderssen
> is an obvious example of someone who had done so recently.
>
> Is this an acceptable state of affairs to you? Do you like being lied to?
> Do you like taking every vendor's "word" as a lie? As a software professional
> you should be ashamed of those who give us a reputation for flashy demos
> and empty promises. I know I am.
Do you or the other 4 or 5 Netscape-baiters on this group have any
evidence to back up your claims of intentional deceit? Are we
supposed to take your word for it?
Ed Kubaitis - e...@uiuc.edu
Jeffrey J. Hoover
Thank you Lloyd!
In article <LJZ.95Ja...@panix.panix.com>, l...@panix.com (Lloyd
Zusman) wrote:
[-snip-]
My opions are my own.
--
________________________________________________________________________
Jeffrey J. Hoover | voice- 408.526.5157
Macintosh Systems Administrator (MacDude) | fax- 408.526.4575
Cisco Systems, Inc. | email- j...@cisco.com
- Help! Help! -- I'm caught in the WorldWideWeb! -
Tom Krishan
: evidence to back up your claims of intentional deceit? Are we
: supposed to take your word for it?
: Ed Kubaitis - e...@uiuc.edu
While I don't consider myself a Netscape-baiter, I will admit to
being personally confused and disappointed by the various Netscape
announcements and final release statement.
From a Netscape Press Release:
------------------------------
> Netscape, Available Now,
> Builds On Tradition of Freeware for the Net
>
> MOUNTAIN VIEW, Calif. (October 13, 1994) -- Netscape Communications
> Corporation today announced that it is offering its newly introduced
> Netscape(TM) network navigator free to users via the Internet. The new
> Internet navigator, developed by the six-month-old Silicon Valley
> company led by Silicon Graphics founder Jim Clark and NCSA Mosaic
> creator Marc Andreessen, is available immediately for free downloading
> by individual, academic and research users.
>
> By making Netscape available free to individuals for personal use, the
> company builds on the tradition of software products for the Internet
> being offered free of charge.
From the Netscape 0.96 Beta:
----------------------------
> Netscape 0.96 (PUBLIC beta version) is free for your personal use,
> subject to the terms detailed in the license agreement accompanying it.
>
> Subject to the timing and results of this beta cycle, Netscape
> Communications will release Netscape 1.0, also available free for
> personal use via the Internet. It will be subject to license terms;
> please review them when and if you obtain Netscape 1.0.
And from the Netscape 0.96 Beta License Agreement:
--------------------------------------------------
> 4. While Netscape intends to distribute a commercial release of the
> Software, Netscape reserves the right at any time not to release a
> commercial release of the Software or, if released, to alter prices,
> features, specifications, capabilities, functions, licensing terms,
> release dates, general availability or other characteristics of the
> commercial release.
So, I can only conclude that Netscape exercised their option of
changing the price of their product for personal use from free to
$39 for release 1.0.
Tom Krishan
kri...@lf.hp.com
steve johnson
methinks that ye of massive paranoia need to subscribe to the ctheory
(conspiracy theory) mailing list.
if you don't like what netscape/marc have done you can do 2 things:
1) don't use netscape
2) write your own client
i suggest both, but especially number 2. it should keep you busy (and
quiet) for a while.
--
GOOD GOD!
- james brown
Nick Gassman
> I am angry at Netscape for not telling the network community that it had to
> change its plans and charge for 1.0, hiding the fact under its 'unlimited
> evaluation for future purchase of the commercial product'; not responding
> to questions directed specifically to Netscape Communications; continuously
> ignoring questions directed to them on the net; and not thanking the people
> who took the time to give bug reports (yes, I did send in at least 3 bug
> reports). I didn't expect to keep getting Netscape for free, but I did
> expect to see some of the problems I reported go away before they started
> charging for their product.
>
Most of the heat is generated about the fact that Netscape said that the program
would be free for certain users and then reneged.
I'm as much annoyed at the fact that they *must* know of a number of bugs such as
Jim and myself have reported, yet there is no one place where they have posted
them. You could argue that a user need only pay for the program once they are
satisfied that it will do the job for that user, but that's not good enough.
When I evaluate a program I do so in light of my current and anticipated needs.
Once I have paid I could then find I need the program to do something that it
ought to do, only to find it won't, and that Netscape knew about it, but didn't
tell me.
I don't think it is reasonable to ask people to pay for a flawed product without
telling them what the problems are first.
--
Nick Gassman - known as ni...@netwiz.demon.co.uk
Interactive Multimedia across The Internet is here... - ask me
Or look at http://www.route-one.co.uk/route-one/