Mailinglist Archive: opensuse (1698 mails)

< Previous Next >
Re: [opensuse] Re: fsh and distros
On 12/21/2011 3:25 AM, lynn wrote:
On 21/12/11 02:51, Brian K. White wrote:
On 12/20/2011 7:50 PM, lynn wrote:
On 12/21/2011 01:25 AM, Joachim Schrod wrote:
lynn wrote:
On 12/21/2011 12:25 AM, Brian K. White wrote:
On 12/19/2011 8:05 PM, Rüdiger Meier wrote:
On Tuesday 20 December 2011, Cristian Rodríguez wrote:
On 19/12/11 17:15, Brian K. White wrote:
We do not WANT that much uniformity among distros.
The lack of uniformity, is a mayor problem, at least from my
developer point of view, it duplicates a lot of work.
Why do we need openSUSE at all? Because it's green?

Uniformity is nice where it does not have too much downsides. I'am
totally fine that there are distros like Ubuntu or Fedora rapidly
throwing away good old unix style to get over some apple/windows
users.
But isn't Fedora/Ubuntu enough? Do we need a greenish mix of them?

What exactly should be the difference between Fedora and openSUSE?
That right there is the $64,000 question.
The consultancy down the road are hiring. It's a deal more than you're
earning now, less work and a company car. You're OK with Yast and
openSUSE. They use Debian.

Job vacancies are beginning to specify not Linux, but a distro.
Well, but that's easy. In that situation, one always knows Debian,
of course. Anybody who knows her or his way around openSUSE and
isn't proficient in Debian within a week (that one needs to be
proficient in any new company anyhow) is not worth hiring. All the
others can tell their future employers that they know Debian,
because, at the time the specific knowledge is needed, they will.

And that's my NSHO as a CEO who does hiring. My company exists
since 1995, and I was always able to distinguish between blenders
that tell me about dists they boasted to be in-and-out and (to take
a recent example) Ex-Solaris admins with 20+ years experience who
were grumpy but knew Linux best practices within a forthnight. No
need to tell who I hired.

Joachim

That's interesting. Do you get them to do anything with a computer when
they attend interview?
L

I don't have to see them do anything on an actual computer.

Just a few minutes of conversation is all. What's required is that _I_
know what _I'm_ talking about, so that when the applicant says
something, I know just how ingenious, or not, that was. A surprisingly
few words can actually indicate so much. One little statement proves all
by itself that the person solves any problem thrown at them by finding
the tools that might do the job, then figuring out how to use the tools,
then successfully using them. Most people just don't do that. They just
know how to do whatever someone else explicitly told them to do.

There is a hands-on proof stage anyways, which is just that we generally
have people do specific jobs as independent contractors for a few jobs
first and then offer to hire them full time. A few stay
quasi-independent anyways by choice and we just always have more work
for them than they can keep up with.

Mmm. That's not what we find. Consultant's here will not work with
Linux. All they do is tell you to replace it with 2008 server and win 7.
Candidates will tell you they can do everything whilst chatting.
Everyone is an expert. Put them in front of a test box with openSUSE and
they are clueless!
L x

I am not talking about having someone say "I'm good at Foo." and believing them. That's stupid.

And what would I care about a consultant who didn't even claim to support linux in the first place if I needed to fill a position centered around programming or sysadminning on linux? I'm not even talking to them in the first place and would never waste my or their time expecting them to do anything on a suse box.

When someone says they're an expert that's generally exactly an example of a few words saying a lot. Real experts don't usually claim to be experts.

When someone says they "...wrote a small bbx program on an old AIX box to run wget to pull down foo web page, pluck out the current prices, and update their bbx application db automatically every 20 minutes 10 years ago..."

Yeah he could be lying, but he'd have to know a lot even just to construct that lie. It's stupid to waste much time worrying about that possibility.

There are so many things going on in that one little statement it's not funny.

But I only know that because I happen to know just how inconvenient it is to work with an "old AIX system" (no nice gnu tools at all, not even bash for the shell), including no wget, so he had to find a wget binary (and any necessary libraries) someone else built (for his particular cpu and version of aix) or compile it himself, and I happen to know what a pain that can be even on a more popular and modern platform let alone something like aix. And I happen to know what bbx is and have written programs in it and know how quirky and inconvenient it is. and I happen to routinely write integration hacks like that, downloading a web page and scraping it for something when there is no nice official api to do it in a better more robust way instead of just living without, and I happen to know how much more difficult this all was 10 years ago when tools were fewer and cruder across the board and apis to access data from web sites was almost unheard of. And on top of all that, no one told him "solve this problem by installing wget and using it to scrape this web page..." no his bosses at the time didn't even realize anything like that was even possible and never occurred to them to even ask, he thought of the possible way to get a great advantage (in their particular business), devised the possible mechanism, found the tools that would do it, then used the tools to implement the plan, to get the final result. Many people can do the final part, implement a plan. But that's almost the useless part. That's the monkey pull this lever part. Fewer people can devise the plan.

When someone says "...so I just installed mediawiki and went through and changed the filepro app so that whenever you press the help key, it just grabs the title of the current screen/menu/db-file/report/whatever and generates a wiki url with that as the page title and fires it off with the terminal emulator escape sequence." Bam! Instant effortless user-generated and/or vendor-generated high quality always up to date documentation for every single place anywhere in the entire app. Ahhh it's so friggn simple and brilliant. Actually it's an ideal example of when engineers use the word "elegant". When you fire off a url to a mediawiki page that doesn't exist, you don't get an error, you just get an empty but perfectly nice page with that title all ready for you to start writing into.
That simple hack made that customers app about 18,000 times better overnight and didn't require very involved coding/hacking of the application, just a relatively simple change added in a lot of places.

That particular job doesn't mean he was necessarily a genius. I do stuff like that all day every day and I know I'm not. It does mean he doesn't need me to figure everything out for him either to get started on some job, nor to deal with problems that arise in the course of it. If in that job above he ended up not being able to get a wget binary, he'd have looked for a curl one, or a lynx one or maybe the box has perl and he could use it's native tcp support and some script he could google up, or he could maybe even find a way to do it with shell scripting and netcat or native tcp support in ksh, etc etc. I know from talking to him in under 5 minutes that that's how he goes about problems in general.

And he absolutely does not call himself an expert at anything.

Conversely, when someone needs me to tell them to use google as a first step...

Conversely if we need an accountant, _I'm_ not the one to interview them. I know nothing about that so they could say anything and it would probably sound ok to me. I'd have to rely on references and perhaps some form of industry recognized test. I wouldn't even know how to tell if they were doing an especially good job after working a while. If we don't get audited or we get audited and come out fine, well that's nice but is that all a good accountant necessarily does? Are there such things as ingenious accounting hacks? Like I don't know improving our insurance deals or something... I have no idea. I could google and read up and I could consult other accountants I happen to know among my friends and family, but the point is I'm sure that a good accountant could talk to someone and pretty much tell if they are a good accountant or not within a few minutes of chatting. If that was the point of the chat, of course not random or empty chatting.

--
bkw
--
To unsubscribe, e-mail: opensuse+unsubscribe@xxxxxxxxxxxx
To contact the owner, e-mail: opensuse+owner@xxxxxxxxxxxx

< Previous Next >