Re: Automated download/update of data files
Daniele Tricoli writes: > On Tuesday 25 August 2015 13:08:12 Ole Streicher wrote: >> This is probably the way to go. However, the original package does not >> update the data on a regular base. It checks whether the data are >> current when they are accessed and downloads a new version if the local >> version is too old. > > What about using the same approach of upstream in the cronjob? If > observations > are <7 days old do nothing otherwise update. This just means downloading every 7 days. The table is updated regularly. Best regards Ole
Re: Automated download/update of data files
On Tuesday 25 August 2015 13:08:12 Ole Streicher wrote: > This is probably the way to go. However, the original package does not > update the data on a regular base. It checks whether the data are > current when they are accessed and downloads a new version if the local > version is too old. What about using the same approach of upstream in the cronjob? If observations are <7 days old do nothing otherwise update. Cheers, -- Daniele Tricoli 'eriol' https://mornie.org signature.asc Description: This is a digitally signed message part.
Re: Automated download/update of data files
Danny Edel writes: > From a system administrator's point of view, the following would be all > right to me: > > * When you install the package, debconf asks if you want the automated > upgrades, preferably telling me how often its going to download them and > information about the source. (And I know I can always change my mind > later with dpkg --reconfigure packagename, although this could be > explicitly stated in the dialog). This is probably the way to go. However, the original package does not update the data on a regular base. It checks whether the data are current when they are accessed and downloads a new version if the local version is too old. Best regards Ole
Re: Automated download/update of data files
m...@linux.it (Marco d'Itri) writes: > On Aug 25, Ole Streicher wrote: >> What is the best way to keep these data up to date in Debian? An >> automated process as written in the pull request [1] is probably not the >> right way, since it is a potential privacy violation. One could let the > I am frankly tired of people screaming "privacy violation" at every > automated process which interacts with the rest of the world. I accept their concerns, even if I don't think that the download of the UTC deviation could cause any privacy concern. There are people who want to have full control how their systems communicate, and they have a valid reason for that. And Debian Policy is to take them serious. So I will. I just seek for a good way to make these people happy and have a good user experience as well. Being tired is not the way to go; solving issues is it. Being constructive on your side would help much. Best regards Ole
Re: Automated download/update of data files
On 25/08/2015 08:44, Ole Streicher wrote: > Hi all, > > What is the best way to keep these data up to date in Debian? An > automated process as written in the pull request [1] is probably not the > right way, since it is a potential privacy violation. One could let the > user manually start the download. However, then he has to keep track of > all these little updates himself (or be educated enough to write > cronjobs, or re-enable the automatic download, which is more that one > could expect from an average scientist). Should one use debconf to ask > the user if he wants automatic updates here? Hi Ole, On the privacy front, I think we probably need to come up with a mechanism with Debian for recording / signaling these issues to the user, so a user can make a decision. As a project, we can look at what we do for users privacy. One principle to follow on cases like this is "keep requests local". Rather than have it broadcast over the whole of the 'net (and any monitored traffic points), request data from as local as possible and make it cachable. If the data comes via a local squid cache, etc. or local CDN then the visibility is much reduced. best regards Alastair > Best regards > > Ole > > [1] https://github.com/astroplanners/astroplan/pull/67 > -- Alastair McKinstry, , , https://diaspora.sceal.ie/u/amckinstry Misentropy: doubting that the Universe is becoming more disordered.
Re: Automated download/update of data files
On 25/08/15 09:44, Ole Streicher wrote: > What is the best way to keep these data up to date in Debian? An > automated process as written in the pull request [1] is probably not the > right way, since it is a potential privacy violation. Hi Ole, I wouldnt say that it' automatically bad to download current data on a regular basis, as long as the system administrator *agreed* to do automated downloads / have the computer talk to the outside world. >From a system administrator's point of view, the following would be all right to me: * When you install the package, debconf asks if you want the automated upgrades, preferably telling me how often its going to download them and information about the source. (And I know I can always change my mind later with dpkg --reconfigure packagename, although this could be explicitly stated in the dialog). Take a look at "dpkg-reconfigure popularity-contest" for an example. * The choice is stored in /etc/packagename.conf or similar. A comment informing that "dpkg-reconfigure packagename" is the easy way to edit the file might be nice. * The package installs a cronjob that does loads the configuration file, and depening on whether auto-downloads are allowed, does the actual downloading/verification/installing of the files or simply does nothing. Since this rests on the system administrator being informed about what the system will do, and giving an explicit OK, I would not consider it a privacy violation. I don't know how complicated this is to implement, but do you think it could be a good design rationale? - Danny
Re: Automated download/update of data files
On Aug 25, Ole Streicher wrote: > What is the best way to keep these data up to date in Debian? An > automated process as written in the pull request [1] is probably not the > right way, since it is a potential privacy violation. One could let the I am frankly tired of people screaming "privacy violation" at every automated process which interacts with the rest of the world. -- ciao, Marco pgpkLEJesvsIi.pgp Description: PGP signature
Automated download/update of data files
Hi all, for astronomy (and probably for other parts of science) we need to access data files that are updated from time to time. An example is the difference between UTC and earth rotation. This data is updated every week and is needed to precisely calculate the positions of stars on the sky [1]. What is the best way to keep these data up to date in Debian? An automated process as written in the pull request [1] is probably not the right way, since it is a potential privacy violation. One could let the user manually start the download. However, then he has to keep track of all these little updates himself (or be educated enough to write cronjobs, or re-enable the automatic download, which is more that one could expect from an average scientist). Should one use debconf to ask the user if he wants automatic updates here? Best regards Ole [1] https://github.com/astroplanners/astroplan/pull/67
