Joseph Galbraith wrote:
> > But I think that's not really the issue.
> > I think that the real issue is that you're hellbent on having the
> > authority over which version of TSVN is installed on my PC.
> >
> > You want me to have the newest version all the time so you get the
> > newest crash reports. And you'd rather that all users live blissfully
> > without knowing that TSVN updates automatically, rather than loosing
> > up-to-date crash reports from those users that decide to opt-out from
> > automatic updating. I think *that*'s the issue.
>
> Well, you might have a half a leg to stand on _if_ TSVN did
> update automatically. But since it doesn't (it simply notifies
> the user that there is a newer version available) I think you
> are at least a ways out in left field here...
Yeah okay, you got me, I used the wrong words.
Errata: "opt-out from the phone home feature".
Stefan Küng wrote:
> > No, because there is absolutely nothing controversial about any of
> > those other options.
>
> Checking for updates is controversial?
> See, here's the point where we completely disagree.
Yup, right.
I find it controversial for any software to call
it's home base without asking me first.
> I consider that feature important
So do I.
I'm not warranting it's removal.
> We don't send *any* data home. So how's that controversial?
As a regular user without the slightest idea of what goes on in TSVN
code, I have no way of knowing that.
Thus I retain that the polite thing to do is to ask me during installation.
> In that case, you could also consider our status cache to be
> controversial because it runs in the background.
Sucking up CPU time and IO bandwidth?
Yeah, you're right :-).
There's a difference however which makes the status cache far less evil.
If you don't like the status cache (even though it's obviously good
for you), you can just uninstall TSVN and you're good.
But if you don't like that TSVN has been phoning home for the last
decade you've used it because you had not discovered the option that
turns this behaviour off, there's no way of undoing what it has done.
You can't poll out the information that's gathered in the log files in
the evil TSVN headquarters (pun intended ;-)), however puny and
insignificant those data may be from your point of view.
> > you've put in the options dialog is in fact not an option at all -
> > TSVN might well have phoned home before the time that you happen to
> > reach the option dialog and find the check box.
>
> If you don't open the settings dialog within seven days after first
> using TSVN, then yes. But I think in that case, it's your own fault.
What what?
If I don't open the settings dialog immediately after installing, and
then meticulously wade through all the options and try to figure out
whether one of them might be an "application bug" aka "phone home
feature", then it's my own fault?
You *must* be kidding me.
> > No other TSVN option has those characteristics, so nobody can
> > realistically complain about those not being in the installer. Nor
> > can I ever imagine that anyone would.
>
> What do you mean with 'characteristics'?
I mean this combination:
* activates without specific user action (not counting 'installing
tsvn' here) and
* transmits (however insignificant amounts of) data to headquarters
> Please don't take this the wrong way. But your completely paranoid.
> [snip]
You misunderstand me.
I'm not afraid of giving TSVN HQ my IP address.
*Obviously* you have that already.
> > You want me to have the newest version all the time so you get the
> > newest crash reports. And you'd rather that all users live blissfully
> > without knowing that TSVN updates automatically, rather than loosing
> > up-to-date crash reports from those users that decide to opt-out from
> > automatic updating. I think *that*'s the issue.
>
> Again, you're assuming that TSVN is evil.
No, I'm arguing that phoning home without my consent is inherently evil.
> Seriously: TSVN does *not* update itself automatically! It only *checks*
> if there's a new version available and then *tells* the user about that.
> The user still can choose not to use the new version.
The point above was that you're afraid that users will never upgrade
if you give them the possibility to opt-out from the "automatic
update" (revised ed. "phone home") features in TortoiseSVN.
The point was *not*, as also explained towards Joseph, that TSVN
updates automatically without asking the user first. I know very well
that it doesn't do that - sorry about that confusion.
> I'd like to correct you, but I can't. I just don't know how I can
> explain to you that a simple check for a new version is not evil.
> Sorry, I really don't get it.
Ok.
But you wouldn't mind if your users had the ability to opt-out from
the phone home feature?
As an example, let's say that it's a check box and it has a tick in it
per default.
The installer will need a description for the check box, so here's a
concrete suggestion:
"This is what we do; we poll to see if a new version is available once a week.
We collect no data from your PC.
The server will log a timestamp, your IP and some HTTP header
information, which is absolutely necessary from a security standpoint.
From that information we can theoretically gather approximately how
many TSVN users there are as well as for approximately how long you've
had TSVN installed, but we don't make a habit of snooping around in
those log files, as there are usually better ways of finding out."
> And of course you think that's the fault of Firefox.
> [snip]
> Yes, very evil indeed.
No, I don't think that's the fault of Firefox.
I just prefer to upgrade less often because it gives me a much more
stable environment.
Even if it is theoretically 0.001% more likely that I'll catch a virus.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tortoisesvn.tigris.org
For additional commands, e-mail: dev-help@tortoisesvn.tigris.org
Received on Wed Apr 12 21:35:34 2006