|
Grex > Coop12 > #57: Proposal: Users shall be able to withdraw their text | |
|
| Author |
Message |
| 25 new of 168 responses total. |
davel
|
|
response 113 of 168:
|
Nov 20 13:05 UTC 2001 |
What Mary said.
|
md
|
|
response 114 of 168:
|
Nov 20 13:28 UTC 2001 |
Something I like to do on M-Net is enter the funniest thing anyone ever
read, and then the instant someone says, "That's the funniest thing I
ever read," I scribble it, which on M-Net means it's gone forever. If
the timing is right, which I've managed to make it be a couple of
times, the response is entered and scribbled in the space of a few
minutes and no one after that gets to see it. Most satisfying.
The Grex version of this trick is to enter the response "Gotcha!" at
some critical point in an item (preferably an item about the censored
log, natch) and then immediately scribble it. Someone goes to read the
censored log to find out what you said, and there's your gotcha.
This sort of playing with the system's peculiarities used to be much
more popular than it is now. Call me immature and I'll call you old.
|
remmers
|
|
response 115 of 168:
|
Nov 20 15:55 UTC 2001 |
This response has been erased.
|
md
|
|
response 116 of 168:
|
Nov 20 20:34 UTC 2001 |
Something else you can do on M-Net is enter an item on one subject,
wait until a bunch of "mnetters" enter responses to it, then edit the
item to something else entirely. I once entered an item announcing
that I was leaving M-Net for good, then after a bunch of people had
responded (incredibly, considering the abuse I heap on them) with "No!
Please don't go!" I edited the header and text to "I'm going to the
movies tonight." There are much better uses you could put this feature
to, obviously, but that's the only time I ever did it. I don't know
why this isn't done more often on M-Net. (Or maybe it is and I haven't
noticed?)
|
dpc
|
|
response 117 of 168:
|
Nov 20 21:37 UTC 2001 |
I am voting "yes" on krj's proposal. Long overdue.
|
davel
|
|
response 118 of 168:
|
Nov 21 14:31 UTC 2001 |
Re 115: heh.
|
polygon
|
|
response 119 of 168:
|
Nov 21 16:24 UTC 2001 |
I have also cast my "yes" vote on this proposal.
|
gelinas
|
|
response 120 of 168:
|
Nov 21 16:41 UTC 2001 |
I'd be surprised if anyone thought I had voted against it. :)
|
jp2
|
|
response 121 of 168:
|
Nov 21 18:32 UTC 2001 |
This response has been erased.
|
albaugh
|
|
response 122 of 168:
|
Nov 22 09:00 UTC 2001 |
Sorry to be an albaugh-come-lately to the discussion of this motion (voting
is already underway), but by testing I dicovered this:
Response not possible. Pass? scribble 99
Cannot censor frozen items!
This may have been well known to many/most grexians (somehow I doubt it), but
it strikes me that closing the censored log would not afford "equal
protection" to those who wish to publicly retract their comments, in the
case of frozen items. In the worst case, a "malicious" item creator might
deliberately freeze his item, just to prevent off-the-cuff Joe from making
a retraction.
It could easily be argued that this is an issue with the scribble command,
and doesn't have a bearing on closing the censored log. But since the vote
extends for another week, I'd be interested to hear what other people think
about this.
|
remmers
|
|
response 123 of 168:
|
Nov 22 18:15 UTC 2001 |
I wasn't aware that frozen items couldn't be scribbled, but I just
tried it out, and that appears to be correct. Hmmm...
|
gelinas
|
|
response 124 of 168:
|
Nov 22 20:36 UTC 2001 |
So in separate action we find a way to scribble frozen items.
|
janc
|
|
response 125 of 168:
|
Nov 22 21:03 UTC 2001 |
I could probably teach backtalk to allow that.
|
janc
|
|
response 126 of 168:
|
Nov 23 05:11 UTC 2001 |
I dug around a bit in backtalk's source and figured a fairly painless way to
enable this. The next backtalk release (if I ever get it out the door -
version 1.1.7 has a huge list of deep changes that haven't all been ironed
out yet) will be configurable to enable erasing responses in frozen items,
if desired.
There are sound technical reasons why this wasn't allowed in Picospan.
Erasing requires editing a flag bit in the header of the response, but frozen
items are marked frozen by having write permissions turned off, making it
impossible to edit the file. My kludge was to thaw the item before opening
it, and then immediately refreezing it after it is open. This means that the
item is very briefly thawed whenever such a operation is executed - or at
least partially thawed - state information is also in sum file, but that is
mostly less authoritative. So enabling this might make it possible for a
really hardworking user to post a response to a frozen item. Which isn't so
horrible a thing but it's kind of an ugly anyway. In the best of all possible
worlds, we'd find some other way than permission bits to flag an item's state.
There are also likely philosophical reasons why people might think it
shouldn't be possible to erase from a frozen item. So whether or not we
enable this Backtalk feature is still open for discussion.
|
kaplan
|
|
response 127 of 168:
|
Nov 25 05:10 UTC 2001 |
I voted against the similar proposal last time around, and I'm voting
against this one too. I would support a proposal to eliminate the
scribble command entirely.
|
devnull
|
|
response 128 of 168:
|
Nov 25 05:25 UTC 2001 |
Re #126: and why isn't there a way to have locking to prevent anything but
the process doing the scribbling from changing the item?
|
janc
|
|
response 129 of 168:
|
Nov 25 05:41 UTC 2001 |
Right now it does
chmod(file, read-writable)
fp= fopen(file,"r+");
chmod(file, only-readable)
and then goes on to do it's reading and writing. The window of vulnerability
is between the two chmods, and it's a pretty short window. I can't lock it
until after I've opened it, but I suppose I could open it read-only, apply
an advisory lock, chmod it, open it read-write, chmod-it and close the
read-only file handle. Hardly seems worth the effort though. Not sure if
it would work with all the different locking libraries we support.
|
jp2
|
|
response 130 of 168:
|
Nov 25 06:15 UTC 2001 |
This response has been erased.
|
cmcgee
|
|
response 131 of 168:
|
Nov 25 16:06 UTC 2001 |
I voted to close the scribble log t everyone but staff. I would vote
against getting rid of the scribble command entirely.
|
janc
|
|
response 132 of 168:
|
Nov 25 17:02 UTC 2001 |
Re #130: because backtalk does not run as superuser.
|
devnull
|
|
response 133 of 168:
|
Nov 25 22:51 UTC 2001 |
Re #132: But you could call a suid program.
|
devnull
|
|
response 134 of 168:
|
Nov 25 22:52 UTC 2001 |
Re #129: I suspect that if you decide you care, the fix might be to have
item frozenness marked by something other than unix permissions.
|
janc
|
|
response 135 of 168:
|
Nov 26 04:20 UTC 2001 |
I agree with the last, except that I have to maintain compatibility with
Picospan, which is essentially inalterable.
|
albaugh
|
|
response 136 of 168:
|
Nov 26 14:57 UTC 2001 |
Is mdw lurking on this, or will/can he weigh in?
|
remmers
|
|
response 137 of 168:
|
Nov 27 17:04 UTC 2001 |
Even if Picospan is unalterable, couldn't one make whatever method
backtalk uses to expurgate/scribble frozen items available to telnet
and dialup users as well? Code it as a standalone suid program.
|