OSDev.org

Hi,
I was just wondering about why some users have stars below their username, while others don't.
Is there any particular reason for that?

Thank you!

That depends on user post count. More posts - more stars.

Hi,

I think this rating system is silly. Having written more posts doesn't mean you are more useful than others who have written less posts. For example, Brendan and (Com)buster deserve their five stars, but Cotton509 (if someone knows/remembers him) doesn't deserve even his two stars. Right now I have two stars too, but I don't really deserve them, one star would more than enough.

Don't trust stars, trust users.

Regards,
glauxosdev

Edit: Shouldn't this be moved?

We probably all agree on that the stars system is bad and doesn't give any useful information, maybe except if someone's new. Upvotes for posts would be a solution. But the issue here is that we'd need a new forum software (already discussed a few months back, not going to happen soon) or to install a phpBB mod (@Mods: can you install phpBB mods? I expect no as the most probable answer, but you never know ...).

Postcounts are a "poor" heuristic because they don't reflect the comparative difference between people. Reputation systems are possibly a worse heuristic because apart from being just a static piece of information, it favours feeders over people just being active.

Stars may not be perfect, but they're a hint in the right direction, and having to use a banned member to provide a counterexample illustrates little more of a point than the fact that there is no "perfect" system, and neither would I have any reason to deny you your two stars

By the way, upvote is worst, as proved by SO, an obvious question in elementary level usually get more upvote than a fruitful discussion between experts, simply due to population distribution.

I'll just ignore the stars instead of introducing another star system.

Hey, I've earned my lucky stars!

They're always after me lucky charms.

glauxosdev wrote:Cotton509

Damn, I've just read these posts, I hope, he was a troll.

There is a correlation between stars and experience so it's not that useless. You normally need to carefully read posts to see who is right or not in a topic. If in doubt, stars supply a hint.

What? Only four stars. I must be wrong.

What about having an option to follow users and 'stars'; that are based on the number of followers--not on the number of posts?

The number of stars mean OSDev addiction intensity level.

Just kidding, the forum did have a discussion about this in an old thread: http://forum.osdev.org/viewtopic.php?f=11&t=26998

Here's the table given in the above link:

• 0025 posts - ★
• 0075 posts - ★★
• 0150 posts - ★★★
• 0300 posts - ★★★★
• 1000 posts - ★★★★★

It's not 1000.
Through site admins should choose a star and update it monthly (to active users) by user knowledge.

It's not just stars but registration date for me. Those two metrics for me combine to be a degree of trust. I know it's still a completely irrational scale.

But my thinking is that the older an active user is compared to how much they have written in that time determines trust. To old and little posts... or recent and alot of posts are at the bottom of my scale. Old but very active or new with few but flawless posts score way higher.

This is not something I do consciously.
Just the result of some self-reflection.

A site I frequent has a 'tweakotine'-metric tho that measures and shows something like this.

It is irrational. What do you want it to rate ? Social behaviour ? Or " osdevel level "?

If you want stars to have a osdev meaning, how about White wolf ascension ones ?
* : Ability to read, reproduce fully guided tutorials and understand their general idea
** : Ability to do small changes and adaptations from existing code to your idea. Your kernel really looks like someone elses.
*** : You need to understand that under your Os, there is a hardware, a firmware, some peripheral. They can be flawed, or designed badly. You know how to practice black magic needed to achieve a goal. You know how to make major but no huge structural change in OSes.
This includes abusing triple fault to reboot, working around CPU/emulator bugs, doing some nop nop between sti and hlt, using recursive paging..

**** : You know that aside your OS, they are many other designs, some are very bizarre. You know how to analyse something really alien. This knowledge allows you to make drastic changes in an OS, create weird ones.

***** : You have realised that conceptually, your idea of kernel has some flaws . You know how to spot them, and you have a feeling how to make your kernel " the real thing ". You are working on them. But you are sad, there is a lot of work, and you don't want to break compatibility. But you are correcting them nevertheless. You start flagging deprecating APIs..
" oops, i shouldn't have imaginated a fork/pthread_cancel method "
With this amount of stars, you are able to correct the reality of some OSes.

****** : You are an OS guru. You leave a something behind you, good people follow you and this feed your wisdom.
You are able to make some idea appear in some OS. You may even have some bots quoting you on IRC, or some people stealing your graphical compositors .
Not only you are able to change your OS, you are also able to subtly influence others.

*******: You are not in the same world as us. Your influence extends to other spheres, like hardware, libc, algorithmic..







Do you know people who match those ? How many stars do you give yourself ?

Around 4.
I don't understand how do you mean analyse something alien.