I'm not in the least a fan of Oracle, but I think this could still be a safe move for a free EL distribution.
It don't see a major drawback here:
- You get faster and more reliable updates, especially important for security updates.
- The release schedule will be more predictable (I know a lot of people moved to ScientificLinux while waiting for the CentOS6 release)
- It's binary compatible with RHEL (and therefor CentOS). Meaning, that you can drop the Oracle repos and move back if you don't like the behavior you're seeing upstream.
I think that's an accurate summary. And I think the point that you can always switch back to CentOS or Scientific Linux is an important one. Ultimately, this is a topic where a lot of people have religious devotion, so for me it's really a question of whether pragmatism outweighs zealotry.
Yes, you're absolutely right, and that wasn't my intent. (If I could edit it, I'd tone it down a little.)
My goal is just to have people consider the case on its merits.
If you've thought about it and concluded "OK, I see what you're offering, but I'm distrustful of Oracle based on past behavior, and that outweighs your benefits", awesome. I might not agree, but at least we had a reasonable discussion about it.
As opposed to, e.g. "Oracle bad. Google good. Microsoft bad." (Which I'm totally guilty of myself, so I get where people are coming from.)
Well, in people's minds it's not only "past" behavior, but also the quite current Android/Oracle/Java argument. These things (just like the Microsoft UEFI topic) are discussions about very fundamental issues that could have impact on the freedoms and livelihoods of people here.
I'm not sure how much success you'll have getting people to separate their issues with the larger entity from considering your project. After all, when it comes to business decisions it is understandable that people would want to take these issues into account.
What I could imagine to happen is your project building its own separate (and good) track record from Oracle. I think some people asking for a reverse-to-original option/script might have come off as insincere, but I think it would be a crucial part. If you can say "try it, use it, and if you don't like it or anything happens, there's an up-to-date and tested way of returning to the original", it could alleviate peoples fears for lock-ins.
But it'll be hard to work against things like peoples fears that one of the reasons is an attack on RedHat. The only way to deal with that will be time and good behavior (not doubting that you'll do that, but you came out of a perceived lions cage. So the gazelles will watch you for a while to make sure you aren't a lion after all).
I compliment you on your enthusiasm, wish you success in bringing Oracle better reputation in the Linux community, and respectfully submit that I am distrustful of Oracle based on past behavior - specific to my experiences, not altruistically to the community in general.
And for zealotry analytics, Oracle:bad Google:bad Microsoft:bad Apple:bad. Just because I use an OS (or several) doesn't mean I completely trust the intents of the companies supplying that OS (or several).
There isn't a reverse script yet, but you're right, making one would be fairly straightforward.
'--reverse' is something I'd like to see implemented, though, if nothing else because it'll assuage people's fears, and makes it feel less risky to experiment.
It don't see a major drawback here:
- You get faster and more reliable updates, especially important for security updates.
- The release schedule will be more predictable (I know a lot of people moved to ScientificLinux while waiting for the CentOS6 release)
- It's binary compatible with RHEL (and therefor CentOS). Meaning, that you can drop the Oracle repos and move back if you don't like the behavior you're seeing upstream.