Red Hat Bugzilla – Bug 433105
[RFE] Try to integrate smolt with The Linux Counter
Last modified: 2009-06-22 16:52:14 EDT
Description of problem:
The Linux Counter , founded by Harald Alvestrand (http://counter.li.org) has a
long tradition of collecting statistics about Linux machines.
The project offers a script, called „machine-update”, that periodically updates
some data (http://counter.li.org/scripts/).
However, this has a few disadvantages:
- it is not automatic (machine-update must be installed and updated manually)
- not all data initially collected by the script is updated automatically
On the other hand, smolt has far better hardware data collection and level of
I think both projects should benefit if the statistics collected are unified in
a common format and a common collecting "script" is distributed to users, in
.rpm, .deb packages.
Please consider unifying and automating the mechanism of collecting statistics
and distribute .rpm packages.
There is one more information that worths to be added here:
hardware4linux.info seems to be Debian's official choice for a hardware profiler, equivalent to smolt - please see:
Even RPM packages for this tool are officially available:
IMHO, integrating as much as possible these three projects (goals, protocols, mechanisms, resources, etc.) and making hardware data collection as automatic as possible will *greatly* enlarge the opportunities to write 100% free drivers for various hardware and, ultimately, increase Linux adoption.
We cannot afford getting trapped in distribution-based "religious wars", since this will only serve the cause of proprietary software.
So PLEASE search for an unification mechanism.
After studying online a lot more about this, I'm able to say there are at least five significant projects with similar goals:
- smolt (https://fedorahosted.org/smolt);
- The Linux Counter (http://counter.li.org), via the machine-update script (http://counter.li.org/scripts/machine-update);
- hardware4linux.info (http://www.hardware4linux.info), which seems to be the preferred tool in Debian/Ubuntu/Kiwi;
- FSF's Linux Inventory Aggregator (http://www.dracoware.com/ppl/rtwomey/inventory.shtml, please see details at http://www.fsf.org/resources/hw).
- Hardinfo (http://hardinfo.berlios.de/HomePage), which seems to have a Gentoo flavour.
IMHO, we need a cross-distribution hardware profiler, as *automatic* as possible and with *maximum details* about the hardware it scans.
We also need a Windows tool to scan computers we *prospect* to run Linux, in order to quickly submit to driver developers as much informatuon as possible about new or "exotic" hardware on the market.
Sorry, I don't want to make it sound like I'm ignoring you but everything you've mentioned has already been thought of and discussed.
Ideas we have plenty of.
What we're missing is coders.
Also, smolt has a ticketing system -
If the owners of those websites thought it was important for smolt to be integrated with them.. They'd have contacted me. You're asking me to do work to benefit them, which is good valuable work, don't get me wrong. But I don't have the time to do it, and if they don't see the value in doing it I'm not sure why I should.
I agree with your dissapointment, but I will not comment that here.
I systematically scan all hardware I have at hand (in my company) with these tools:
- LinuxCounter (if there's a machine permanently running Linux)
- GNU's Linux Aggregator
- lspci -vvv
and I submit the complete data to email@example.com (please see http://www.fsf.org/resources/hw)
Not being a coder myself, this my (very modest) way to contribute to the general effort: provide free driver developers with as much hardware information as possible. I have a strong conviction that good quality free drivers are esential in Linux mass adoption, especially in the desktop segment.
Please note that I'm not only scanning machines that actually run Linux in my company's network (routers, servers), but any computer that I suspect having more "exotic" hardware, etc.
In this respect, having a "pocket tool" that will gain as much hardware information as possible, in an automatic way, would be an immense benefit.