Broadband Update Bigotry

On Monday, April 25, 2005, at 03:34 PM, Sander Tekelenburg wrote:

> At 08:38 -0700 UTC, on 2005/04/25, Marilyn Matty wrote:
>
> [...]
>
> [MacAddict]
>
>> It isn't free - the developers that participate on the CDs pay to be
>> there.
>
> Not necessarily all of them. MacAddict also actively seeks out
> developers
> asking permssions to distribute their software in the MacAddict CD. I
> know
> for a fact, because that's how my shareware[*] ended up on their CD.
> Twice.
> (In fact, that's how I learned about MacAddict :) I'm not a magazine
> person.)

It would be rather like public service ads you see in newspapers and
magazines; if there's leftover space, it's the best way to fill it.

>
> As for them carrying Apple's system updates, I don't know if they'd
> charge
> Apple (a lot) for that. I would imagine that, if many Apple users have
> no
> broadband available, it would help them sell their magazine. (I may
> well be
> wrong. I don't know the magazine business.)

It's too expensive, and if you're going to give something away for
free, you don't give it to someone with the wherewithal to become a big
advertiser. You'd be sacrificing all hopes of a mega-contract. What
they could do, as I mentioned, is give some slots in exchange for
mega-bucks.

But Apple is unlikely to spend huge amounts on an issue by issue basis
in Mac magazines. People who read Mac publications are already
convinced the products are good and are probably devoted buyers.
They're likely to upgrade and buy new equipment without the prompting
of advertising, and the editorial will be covering it anyway.

>
> But even if it would cost Apple, it seems to me it could be worth it.
> After
> all, we're talking about updates that are in large part bug fixes. It
> seems
> to me it would be wise for Apple to do their best to get those bug
> fixes to
> their customers. A happy customer will come back when it is time for
> something new. Leaving people messing with .0 system versions is less
> likely
> to do that.

I think the problem with this is timing as well as cost. It takes weeks
and weeks between releasing the bug fix and the CD ending up in the
hands of a customer. By the time the bug fix is released, there could
be one or to more that have been issued right behind it. This is
particularly true for newly released system software.

>
> OTOH, it might be that Mac users more often have broadband available. I
> wouldn't be surprised if that were the case. But you'd have to /know/
> what
> percentage of Mac OS X Panther/Tiger users are stuck without broadband
> in
> order to know what the effort is worth.
>

I suspect that broadband users represent a higher % of users who
upgrade routinely.

>
> [*] Shameless plug: <http://www.euronet.nl/~tekelenb/software/> :)

This stuff sounds really cool!

Marilyn

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 23.4.2005, space aliens observed Curtis Wilcox saying:
>>Do you really believe that somebody distributing malware will print
>>his own CDs containing the malware (which have to look like Apple's)
>>and then somehow replace a shipment of real CDs with the fake ones,
>>especially if he can just send his malware via e-Mail and reach many
>>more gullible people who will promptly execute the attached file?
>It doesn't have to be intentional. Some of those magazine CD-ROMs in
>the 90s ended up with viruses on them simply because the computers
>used to create them were infected.

The chance of that happening with a CD is just as big as it happening
with the online download, and if it happens to the CD, you can at least
issue a recall if you notice it soon enough.


And John Welch wrote:
>No, you find someone shipping updates via their own CDs, (user
>groups, etc.) and pervert them. Or you do what the MS phishers do,
>and use a combination of spam and bad downloads.

They can already do that right now. How would officially including
updates on magazine CDs change that?

lucas

- --
"Success is not to be pursued; it is to be attracted by the person we become."
  -- Jim Rohn

-----BEGIN PGP SIGNATURE-----
Version: PGP SDK 3.2.2

iQA/AwUBQm6QF7XYdom/dB2cEQJTKwCgpYbVv4YfRbX6fNKP34EuocMPkm8AoINM
t6kEqZ58qYVaIqMDungY/xSE
=AOS6
-----END PGP SIGNATURE-----

Sander Tekelenburg wrote:

>As for them carrying Apple's system updates, I don't know if they'd charge
>Apple (a lot) for that. I would imagine that, if many Apple users have no
>broadband available, it would help them sell their magazine. (I may well be
>wrong. I don't know the magazine business.)

Several years ago, Apple made it clear that it doesn't want any
magazine distributing its updates on CD. This is not an
Apple-won't-pay scenario (the editorial content of that CD is
entirely uncompensated)... this is an Apple-wants-to-control-updates
issue.

-jason
--
Jason Snell: jsnellintertext.com - AIM/iChat: mw jsnell
www.teevee.org - sports.intertext.com

On 4/26/05 14:59, "Lucas K. Mathis" <tidbitslkmc.ch> wrote:

> On 23.4.2005, space aliens observed Curtis Wilcox saying:
>>> Do you really believe that somebody distributing malware will print
>>> his own CDs containing the malware (which have to look like Apple's)
>>> and then somehow replace a shipment of real CDs with the fake ones,
>>> especially if he can just send his malware via e-Mail and reach many
>>> more gullible people who will promptly execute the attached file?
>> It doesn't have to be intentional. Some of those magazine CD-ROMs in
>> the 90s ended up with viruses on them simply because the computers
>> used to create them were infected.
>
> The chance of that happening with a CD is just as big as it happening
> with the online download, and if it happens to the CD, you can at least
> issue a recall if you notice it soon enough.

However, the money lost is MUCH smaller, since replacing a bad online file
takes seconds and no money other than bandwidth charges, that you have a
good idea about anyway.

CDs have other costs that are effectively doubled in this scenario. These
are not minor amounts of money.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

On 4/26/2005 6:56, "Tony Meyer" <ta-meyerihug.co.nz> wrote:

> (The solution, of course, would be to have some sort of checksum, as
> described elsewhere in this thread. But I think P2P has (somewhat unfairly)
> gotten too much of a bad name for big players like Apple or Microsoft to use
> it, at least at the moment).

Not just Microsoft and Apple. This humble human has no intention of using
BitTorrent (or the other things). BitTorrent is a good idea, but assumes
that the Internet is a friendly place.

  --John

At 23:14 -0700 UTC, on 2005/04/26, John W. Baxter wrote:

> BitTorrent is a good idea, but assumes that the Internet is a friendly place.

So does email ;) Yet 99% of its users still sends their mail unencrypted and
happily provide their passphrase several times a day to insecurely fetch
their mail.

Although BitTorrent shares upload bandwidth amongst downloaders (extremely
cool[*]), it still requires a Web server and "tracker" to make something
available initially. You can take that original down after a while and allow
clients to continue to offer the file, but it'll still be the same file.
Unless there turns out to be a design mistake in the protocol, or an
implementation mistake in a client, the only way to distribute a 'bad' file
via BitTorrent thus appears to be for the bad guy to initiate it himself. So
it's simply up to you to decide whether or not you have a reason to trust the
original source - which is no different from all those files people download
from sites over plain http or ftp. BitTorrent has a built-in hash mechanism,
so it shouldn't be possible for someone to insert a malicious variant into
someone else's torrent. (Well, everything can be cracked of course. But for
the moment I'd think Web servers are probably a more interesting target.
There are *many* more of them, and people tend to just click on anything
shiny, not bothering much to verify sources.)

The BitTorrent protocol is open, the client is open source. See
<http://www.bittorrent.com/>. So if there are security issues, peer review
should (in theory) make sure they'll be discovered.

My impression is that BitTorrent has a lot of potential. Consider for
instance musicians (real ones, not those on MTV) offering their music through
the Net. Better than enslavement by the old-world distribution mafia, but
those audio files can quickly run up your site's traffic bill. If BitTorrent
gets as common as, say, browser plug-ins, they could offer their files that
way, sharing the cost of traffic with their listeners.


[*] So cool that, as I recently read somewhere (slashdot?), some people high
up in the movie industry supposedly are considering using it to distribute
movies...


--
Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/>

At 08:04 -0700 4/27/05, Sander Tekelenburg opined:
>[*] So cool that, as I recently read somewhere (slashdot?), some people high
>up in the movie industry supposedly are considering using it to distribute
>movies...

Probably "I, Cringely":

http://www.pbs.org/cringely/pulpit/pulpit20050421.html

--
Ron Risley
www.risley.net

At 6:56 AM -0700 4/26/05, Tony Meyer wrote:
>I can understand Microsoft being wary of P2P distribution, since they are so
>often the targets of malware posing as legitimate Microsoft updates. While
>Apple would have less of a problem with that, I doubt they'd be quick to use
>P2P to distribute updates, either.

Another solution is that Apple/Microsoft could just cut out the
middleman and host the Torrent themselves. My understanding is that
how torrents work the distributor controls the content of the file..

Nick

At 8:48 AM -0700 4/23/05, Lucas K. Mathis wrote:
>The reason why minor Mac OS updated aren't available in grocery stores
>is probably that grocery stores aren't interested in wasting shelf space
>for them.

I'd gander that AOL, Microsoft, etc pay grocery stores slotting fees
for the privilege of being in the store. (i.e. Slotting fees) Given
that Mac users are a small fraction of the population it just doesn't
make sense to pay a fee and duplication for distribution.

Nick

Quoting Nicholas Barnard <nickinmff.net>:

> Another solution is that Apple/Microsoft could just cut out the
> middleman and host the Torrent themselves. My understanding is that
> how torrents work the distributor controls the content of the file..

Bittorrent's help distributors (by offloading some of the their
bandwidth needs
to volunteers) and broadband users. They don't really help analog modem users.

Bittorrent works by breaking the file up into chunks. Each chunk is then
downloaded seperately (and in random order) from somebody. Once a chunk is
downloaded locally, it is made available as an upload to others bittorrent
users.

So Apple hosts the master file, and the torrent file. The torrent file
contains
hashses for each of the chunks, and the file as a whole.

Person #1 begins downloading, she gets torrent file with hashes from
Apple. She
then begins downloading chunk X from Apple.

Person #2 begins downloading, he gets torrent file with hashes from Apple. He
then begins downloading Chunk Y from Apple.

Meanwhile #1 has finished Chunk X and begins downloading Chunk A,
making Chunk X
available for upload (after it passes hash check against the torrent file).

Person #3 begins downloading, she gets torrent file with hashes from
Apple, She
then begins downloading Chunk X from Person #1.

Person #2 finishes Chunk Y, starts on Chunk B from Apple.

Person #1 finishes Chunk A, starts on Chunk Y from Person #2.

Person #3 finishes Chunk X, starts on Chunk A from Person #1

The more people involved in a download (i.e. the more popular it is)
the larger
the available download sites and the more bandwidth savings for the
file owner.

Because each chunk is checked for accuracy at download time, against a
hash list
from the master server, the ability for a rogue client to introduce a
bogus file
(or even just a bogus chunk) is very hard. It's probably easier to break into
the master server and upload a modified file, and a modified hash list.

But bittorrent doesn't make downloads faster for modem users -- they're still
stuck at 28.8 or whatever their speeds are.

Kevin

On 5/4/2005 12:40, "kevinvanhaaren.net" <kevinvanhaaren.net> wrote:

> But bittorrent doesn't make downloads faster for modem users -- they're still
> stuck at 28.8 or whatever their speeds are.

But from your description, bittorrent would help those dialup users with
session limits to get files larger than the session time limit allows.

  --John (still planning to have nothing to do with bittorrent, but others
obviously differ)

On Wednesday, May 4, 2005, at 01:26 PM, Nicholas Barnard wrote:

> At 8:48 AM -0700 4/23/05, Lucas K. Mathis wrote:
>> The reason why minor Mac OS updated aren't available in grocery stores
>> is probably that grocery stores aren't interested in wasting shelf
>> space
>> for them.
>
> I'd gander that AOL, Microsoft, etc pay grocery stores slotting fees
> for the privilege of being in the store. (i.e. Slotting fees) Given
> that Mac users are a small fraction of the population it just doesn't
> make sense to pay a fee and duplication for distribution.
>

I worked for magazines for many years, and it was fairly common
practice to buy more magazine facings at supermarkets than necessary to
minimize or eliminate the presence of competitors.

Marilyn

> I worked for magazines for many years, and it was fairly common
> practice to buy more magazine facings at supermarkets than necessary to
> minimize or eliminate the presence of competitors.

True, but magazines have appeal to many people and they stand alone on
their own. Lets assume that Macs are 5% of the home based computer
population.
Then lets also assume that Mac users have broadband at a 60% rate. Of the
40% who don't have broadband at home only a third of those actually would
even consider finding a CD at a grocery store for an update. By this alone
we've reduced the population who would want grocery store distribution to
0.66% Apple cannot justify this cost.

People who are not on broadband who want an update will also seek other
methods of finding the update. (Once when I was on an internship and
dialup I went to a hotel told a little fib and got free broadband.) (I
consider assisting the hotel IT staff in finding out that the cleaning
staff was disconecting the ethernet cables as payment enough.))

Nick

On 5/5/05 11:58, "John W. Baxter" <jwblistolympus.net> wrote:

>> But bittorrent doesn't make downloads faster for modem users -- they're still
>> stuck at 28.8 or whatever their speeds are.
>
> But from your description, bittorrent would help those dialup users with
> session limits to get files larger than the session time limit allows.

No, you're still stuck with 28.8 pipes. What bittorrent makes up for is the
aggregate speed of the source, so if one source is really slow, the other,
faster sources make up for it.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

Quoting "John W. Baxter" <jwblistolympus.net>:

> On 5/4/2005 12:40, "kevinvanhaaren.net" <kevinvanhaaren.net> wrote:
>
>> But bittorrent doesn't make downloads faster for modem users --
>> they're still
>> stuck at 28.8 or whatever their speeds are.
>
> But from your description, bittorrent would help those dialup users with
> session limits to get files larger than the session time limit allows.
>
> --John (still planning to have nothing to do with bittorrent, but others
> obviously differ)

Yes it would help with that. Bittorrent clients can resume a download at
anytime, it only retrieves missing chunks from the file (and makes all
currently downloaded chunks available for upload.) It's not unusual
for people
(well me at least) to "resume" a download of a completed file. That makes the
whole thing available to other users.

I'm not a big bittorrent user, but when I do I usually try to keep a download
running until I've uploaded at least double the file size back to the other
users. I do cap my upload rate to about 25% of my max bandwidth -- just so it
doesn't make internet usage on other computers impossible.

Kevin

On 5/5/2005 13:10, "John C. Welch" <jwelchbynkii.com> wrote:
> No, you're still stuck with 28.8 pipes. What bittorrent makes up for is the
> aggregate speed of the source, so if one source is really slow, the other,
> faster sources make up for it.
>
I agree that the speed remains slow. But the chunking should help deal with
the session length timeouts (just as Microsoft provided for chunked dialup
downloads of XP SP2).

By session timeout, I mean, for example, that after 4 hours (common value,
not the only one) the connection ends regardless of what is happening.

  --John

On Thursday, May 5, 2005, at 01:28 PM, Nicholas Barnard wrote:

>> On Wednesday, May 4, 2005, at 01:26 PM, Nicholas Barnard wrote:
>>> At 8:48 AM -0700 4/23/05, Lucas K. Mathis wrote:
>>>> grocery stores aren't interested in wasting shelf space for them.
>>> I'd gander that AOL, Microsoft, etc pay grocery stores slotting fees
>>> for the privilege of being in the store. (i.e. Slotting fees) Given
>>> that Mac users are a small fraction of the population it just doesn't
>>> make sense to pay a fee and duplication for distribution.
>>
>> I worked for magazines for many years, and it was fairly common
>> practice to buy more magazine facings at supermarkets than necessary
>> to
>> minimize or eliminate the presence of competitors.
>
> True, but magazines have appeal to many people and they stand alone on
> their own. Lets assume that Macs are 5% of the home based computer
> population.
> Then lets also assume that Mac users have broadband at a 60% rate. Of
> the
> 40% who don't have broadband at home only a third of those actually
> would
> even consider finding a CD at a grocery store for an update. By this
> alone
> we've reduced the population who would want grocery store distribution
> to
> 0.66% Apple cannot justify this cost.

I should have made myself clearer - I didn't mean to imply that Apple
should sell CDs at a grocery store, but rather that premium rack
positioning can be used effectively to sell in and sell through at
computer stores other than Apple's - like the checkout counters of
CompUSA. It might be worth it for them to buy some extra rack space not
only to entice Mac owners to pick up a copy of Tiger, but to convince
PC people that this is an extremely important and exciting new reason
to buy a Mac.

I worked for TV guide about 10 years ago, and at the time they were
still monopolizing the checkout counters at over 90% of US supermarkets
and groceries to make impulse sales as well as to convince consumers
that they were America's source for television information and a
news/entertainment vehicle in aside from program listings.

>
> People who are not on broadband who want an update will also seek other
> methods of finding the update. (Once when I was on an internship and
> dialup I went to a hotel told a little fib and got free broadband.)

It's true, but the easier and more compelling you make the acquisition,
esp. for people who own Macs that are not at all nerdy, the more it
will sell. And in addition to selling, you're reinforcing that Macs are
the coolest, most up to date and exciting system, to switchers and
users alike.

> (I
> consider assisting the hotel IT staff in finding out that the cleaning
> staff was disconecting the ethernet cables as payment enough.))

Smart thinking!

Marilyn

On 5/5/05 16:54, "John W. Baxter" <jwblistolympus.net> wrote:
> I agree that the speed remains slow. But the chunking should help deal with
> the session length timeouts (just as Microsoft provided for chunked dialup
> downloads of XP SP2).
>
> By session timeout, I mean, for example, that after 4 hours (common value,
> not the only one) the connection ends regardless of what is happening.

I think the resume feature is a bit more useful there. But if you have a
28.8 modem, then that's all the data you'll ever get, 28.8Kbps or so. You
can't change physics in software.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

(This is getting remarkably off-topic, so I don't know if Adam/Andrew will
let this through, but anyway...)

> But if
> you have a 28.8 modem, then that's all the data you'll ever
> get, 28.8Kbps or so. You can't change physics in software.

The point with speed is that although the size of the pipe might not change,
you *might* be able to get a faster transfer if you can get the amount of
data transferred closer to the size of the pipe. If I want to download a
file from the US, maybe all I can get is 10Kpbs, even though my pipe is
28.8Kps, but if I download the same file from here in NZ, *maybe*, I can get
20Kbps. My download speed has doubled, even though my pipe is the same.

Of course, such gains are by no means guaranteed, and with P2P I need the
chap in NZ to have a good outgoing connection to supply my 20Kbps.

=Tony.Meyer

On 5/6/05 13:05, "Tony Meyer" <ta-meyerihug.co.nz> wrote:

>> But if
>> you have a 28.8 modem, then that's all the data you'll ever
>> get, 28.8Kbps or so. You can't change physics in software.
>
> The point with speed is that although the size of the pipe might not change,
> you *might* be able to get a faster transfer if you can get the amount of
> data transferred closer to the size of the pipe. If I want to download a
> file from the US, maybe all I can get is 10Kpbs, even though my pipe is
> 28.8Kps, but if I download the same file from here in NZ, *maybe*, I can get
> 20Kbps. My download speed has doubled, even though my pipe is the same.
>
> Of course, such gains are by no means guaranteed, and with P2P I need the
> chap in NZ to have a good outgoing connection to supply my 20Kbps.

That's due to the aggregation of many sources maxing out usable bandwidth,
and is hardly unique to Bittorent. Limewire can do this too. By combining
multiple sources, achieving saturation on your end is easier, even though
each individual source may be slower than your max download speed.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com