Comment by pmoriarty | Hacker News Reader

Comment by pmoriarty | Hacker News Reader

Parent Full thread

"the current company's inertia dominates whoever they acquire and the internal sales-driven culture in developed markets supersedes anything resembling engineering related from a company they acquire"

What engineering-related innovation has RedHat come up with?

As a Linux sysadmin I've noticed that all of the companies I've worked at that have used RedHat chose it not because it was innovative but because:

1 - they could get a support contract

2 - because RH dominated the commercial Linux space

3 - because RH was the OS the admins knew and both wanted to leverage that knowledge and keep their knowledge current for continuing work with this market leader in future jobs (as that was likely the OS they'd use in most jobs to come)

Most of the "innovations" I can attribute to RH have been widely seen in the sysadmin community as negative. Innovations such as systemd, Network Manager, and selinux. The latter two we've often turned off because they caused so many problems. Unfortunately, systemd was too integrated in to the OS to be able to turn off.

There's RPM, but that was created decades ago when RH was not the company it was now, and it's just one packaging format out of many, and not one that IBM even needs to spend billions to buy. The only other positive tech that RH owns that I can think of is Ansible. But they didn't create it. They bought it.

What RH-originated innovations should we impressed by?

The engineering culture is not solely determined by the innovations the company has come up with, so I don't really see the point of the question you ask, with regards to the statement you quoted.

When you say that "most of the innovations [...] have been widely seen in the sysadmin community as negative", well, that can only be true in your own sysadmin community. If you replace "innovation" by "contribution", and make the list include everything RH works on (SELinux, NM and systemd are really a tiny part of it), the sentence becomes factually incorrect, even in your own sysadmin community (I hope).

Also, systemd was not created by RH (LP was not employed by RH at the time), and SELinux is from the NSA's TRUSIX. Labeling SELinux as a "negative innovation" is very weird (not that whole sentence isn't).

When you say that RH's only "positive tech" is Ansible, but that it was an acquisition, that's true for mostly everything RH participates in (Ansible, but also systemd, OpenStack, Ceph, CoreOS, GlusterFS, KVM, LVM/DM, gcc, gdb, systemd etc... were acquisitions or acqui-hires), so I really don't see your point. Does the fact that all of this software originated from somewhere else than RH negates the existence their engineering culture ?

EDIT: made quotes inline

If the bulk of RH's tech was acquired and did not originate with them then why is their engineering culture even relevant? What engineers work by acquisition? That doesn't sound like engineering to me but business.

Now, if you're talking about the engineering cultures of the companies RH acquired, if RH was able to preserve the engineering cultures of the acquisitions they made, why can't IBM?

Also, if by "engineering culture" you're not actually talking about any innovative products that RH itself came up with, but rather the contributions that RH engineers made to open source projects like gcc or the Linux kernel, IBM has done a lot of that too. I'm not sure why that wouldn't continue under IBM's stewardship of RH.

I feel bad for having inserted myself into this discussion. I think I'm actually agreeing with you here, in that there's nothing that shows that IBM couldn't preserve the culture of its acquisition. I just wanted to comment on what I thought was a statement on their engineering culture.

Personally, I just hope that RH can preserve its engineering culture. I can't say anything on whether it will happen or not, because I don't know.

systemd was created when Lennart was a Red Hat employee, for the record.

I stand corrected. There is something wrong in Wikipedia, as he's cited as working for Red Hat since 2011, while systemd was created in 2010.

I suppose that nobody could find another source than him saying that he works for Red Hat in a FOSDEM 2011 interview.

If you have such a source, I suggest correcting Wikipedia, as it may induce others in error.

https://lwn.net/Articles/299211/ is from 2008. This is the first pre-systemd source I could find, but PulseAudio became the default in Fedora about a year before so he might have been employed at that time, too.

> The only other positive tech that RH owns that I can think of is Ansible. But they didn't create it. They bought it.

Ansible was created within the Fedora Project (owned by RH) as "func" originally. The creators left RH to found AnsibleWorks, which was then acquired by Red Hat in 2015.

SELinux is not purely Red Hat innovation, it largely came from the NSA's Trusted Systems Research Group[0]. That said, you shouldn't be turning off SELinux. Its like saying you should just chmod 777 all your files as you otherwise cannot figure out why your web server won't host your PHP script. On a basic level, it really doesn't take too long to figure out how it works and how to get whatever you're doing play nice with type enforcement.

Perhaps the surface area business value proposition to paying customers is the incorrect place to look for innovation? RedHat has done a tremendous amount for the Linux kernel over decades, and Ansible was basically a RedHat employee project that was made in response to frustrations and spun out and came back into its fold. Also, prior to the advent of yum (RIP Seth Vidal) rpm was all we had. A lot of "innovation" isn't even hardcore low-level engineering as much as standardization and stewardship - none of these things IBM has been known for since the late 80s perhaps. Selinux was also originally an NSA research project that was adopted quickly by RedHat because most of their customers are security-focused (at least on paper).

Not going to disagree that a lot of what RedHat has championed recently has been controversial. What kind of innovation isn't controversial though until usually many years afterward?

Also, cultures can be relative to each other and are not definable in single dimensions. But nobody besides the most Kool Aid soaked of IBMers could say that IBM is more of an engineering-centric culture than RedHat.

Reply on news.ycombinator.com