Is there any accurate way to rank distros?
Distrowatch has its "page hit rankings". The trouble with that is some distros have click happy legions who nullify any hint of accuracy.
I've tried to think of a better way but I'm not doing too well with it.
1-What if only the owner of a distro was allow to vote or rank distros? Ask them to rank the top ten distros and any more they wish to rank. I guess there would be a temptation to rank their own distro first but that may be ok. Peer ranking may get us closer to accuracy. We need this, but it may be a lost cause.
2-Find some way to Fix what Distrowatch does. Weighted average, standard deviation, etc.
3-So what about LQ's own attempt to do something like this. Members Choice Awards. It is what it is but nothing is going to be perfect.
This is just my amateurish opinion so please don't take it too seriously. I'm just one fish in the sea.
I've tried to think of a better way but I'm not doing too well with it.
1-What if only the owner of a distro was allow to vote or rank distros? Ask them to rank the top ten distros and any more they wish to rank. I guess there would be a temptation to rank their own distro first but that may be ok. Peer ranking may get us closer to accuracy. We need this, but it may be a lost cause.
2-Find some way to Fix what Distrowatch does. Weighted average, standard deviation, etc.
3-So what about LQ's own attempt to do something like this. Members Choice Awards. It is what it is but nothing is going to be perfect.
This is just my amateurish opinion so please don't take it too seriously. I'm just one fish in the sea.
Total Comments 3
Comments
-
If you figure out a way to remove response bias (any bias, honestly) you would likely win some awards and worldwide recognition. I'm going to go out on a limb and say that since no one has solved this one yet, odds are pretty good we'll be dealing with bias for many, many years to come.
Not to say the ideas presented are without merit, of course =) Your first duty would be to define "ranking" very precisely so you have some hope of achieving a goal instead of floundering about some vague concept of "better".Posted 04-24-2016 at 12:13 PM by rocket357 -
First, great post, because it provokes thought. Thank you.
In addition to what rocker357 said, there is also this:
Distros--well, computer OS's--can be used for so many different purposes that I think a single ranking scale is whistling in the wind.
There are not just server and home use.
There are business servers, web servers, and file servers. They are all servers, but they serve(r) different purposes and would be evaluated using different criteria. There is also office use (documents and general business), office use (accounting), office use (engineering and design), audio-visual and/or graphic creation and editing, general home use (email, web, some documents, photo editing), and those are just the ones off the top of my head. Each usage would necessitate a different weighting of metrics.
And, frankly, it's not so much the distro that matters as the application suite, because every distro still has the Linux kernel at its core.
Distrowatch's method is sort of a popularity poll and doesn't pretend to be anything more. It's meaningless fun, like a carny game.
Just my two cents.Posted 04-24-2016 at 09:08 PM by frankbell -
Posted 04-24-2016 at 09:52 PM by rocket357