I've been banned from the Marshall Forum...

Not sure I'd call Lennart the devil, but close enough. He kinda made up for a little bit of that trainwreck with Pulseaudio, I guess. Not a fan myself (JACK ftw) but if you want surround playback on Linux that's pretty much the only way.

960px-Lennart_Poettering.jpg




Most people don't expose any ports in their NAT anyway using IPv4. Worst thing they can do is attempt a DoS (flooding/swamping your uplink), which is unlikely to work without a little diversification of source IP's (botnet, etc), pretty much. The old "I've got your IP now" adage is a bit 90's, really.

I expose some services, have done for decades. I also run IDS and IPS, have backup uplinks, etc. I see people trying things from time to time, but they've never managed much more than slow down my nightly logrotate jobs.

Cheers P,
I was quite pleased, months (or years) ago, when you said that you were using the old school gear - I mean Jack.
That is my preference too.
A few years back I had a Ubuntu system with System D and Jack would frequently run away and spam the logs.
This never happened with SysV.

I deal with these difficulties by mobilising simple rules such as...
... never use System D
... never give in to Pottering
... never engage with Ubuntu nor any of the spawn from it's womb.
 
And on a third hand I think I have heard you play nicely.

Oi, where are you getting all these hands from?!

I deal with these difficulties by mobilising simple rules such as...
... never use System D
... never give in to Pottering
... never engage with Ubuntu nor any of the spawn from it's womb.

Systemd is fairly reliable by now, has been for a while. Various devs did quite a lot of mass fixing and "unducking" of its shortcomings. Binary logs and so on are inherently still a part of it, but at least it plays nice with most things and can be tamed to a degree. I dislike yet use it on Debian for practically all purposes except embedded stuff.
Poettering gets a lot of criticism and justifiably so, but so should the devs who approved and integrated his work. An often overlooked angle is that Poettering and Torvalds had some rather heated exchanges because of, amongst other things, the quality of Lennart's code. Linus went so far as to make other devs aware of this and basically block his contributions until the quality improved. If you take a look at Mitre you can get an impression of just how many CVE's he had originating from his code. In revenge he then broke a lot of systems and operations by commandeering the `debug` flag for the kernel, using it for systemd instead. That's bordering on sabotage. There's more, too.
I share the overall dislike for Ubuntu, yet ultimately it's just a more up to date Debian with a lot of bloat and the option of running a pre-compiled low-latency kernel out of the box. The latter is sort of nice.

I was quite pleased, months (or years) ago, when you said that you were using the old school gear - I mean Jack.

Jack works like an audio engineer would expect. Pulse works like a C developer would expect. OSS on Linux works like an 80's UNIX programmer would expect. ALSA sits back and laughs.

Quite honestly, Pulse is not bad for modern desktop integration. OSS did need replacing and ALSA alone wasn't really up to the task, thinking ahead. Having Pulse and Jack play nice together is a bit of a balancing act but stable when done right.

Arguably it *might* have been better for old Lennart to have invested more in Pulse, replaced the whole Linux audio stack including ALSA as the base (but retaining compatibility by making it a layer on top, co-existing with the rest as far as possible) and left systemd as a dusty repo somewhere. Don't get me wrong, I have nothing against ALSA, but I'd rather fight a monolithic audio stack than a monolyth which controls my kernel args, services, boot process, system logging and more.
 
Back
Top