| Time |
S |
Nick |
Message |
| 01:06 |
|
|
aditsu joined #sourcefu |
| 01:25 |
|
prologic |
haha |
| 01:25 |
|
prologic |
yeah |
| 04:30 |
|
|
aditsu_ joined #sourcefu |
| 09:57 |
|
pdurbin |
prologic: http://blog.octo.com/en/openshift-3-private-paas-with-docker/ |
| 10:13 |
|
prologic |
https://github.com/openshift/origin |
| 10:13 |
|
prologic |
http://www.openshift.org/ |
| 10:13 |
|
prologic |
veryinteresting |
| 10:13 |
|
prologic |
:) |
| 10:18 |
|
* prologic |
spins it up |
| 10:28 |
|
prologic |
hmm nope |
| 10:28 |
|
prologic |
few bugs still I guess |
| 10:28 |
|
prologic |
I'll try it again later |
| 10:29 |
|
prologic |
I wrote up a quick docker-compose.yml |
| 10:29 |
|
prologic |
you gotta admit though autodock-paas is kind of a lot simpler to both get going and use |
| 10:29 |
|
prologic |
:) |
| 10:32 |
|
prologic |
https://github.com/openshift/origin/issues/3072 |
| 10:35 |
|
prologic |
it *is* pretty impressive though |
| 10:35 |
|
prologic |
I may actually consider running up a clsuter myself |
| 10:45 |
|
pdurbin |
ah, cool. you left a comment |
| 10:59 |
|
prologic |
I did |
| 12:31 |
|
|
aditsu joined #sourcefu |
| 13:57 |
|
pdurbin |
for anyone reading these logs in the future: https://github.com/prologic/autodock-paas :) |
| 14:07 |
|
|
bear joined #sourcefu |
| 22:25 |
|
prologic |
thanks :) |
| 22:27 |
|
pdurbin |
prologic: sure. I was singing your praises today: http://irclog.iq.harvard.edu/dataverse/2015-07-29#i_22107 |
| 22:37 |
|
prologic |
ahh |
| 22:37 |
|
prologic |
on the contrary containers talking to each other is something I do all the time |
| 22:41 |
|
pdurbin |
so you fear not the dragons at http://martinfowler.com/bliki/images/microservice-verdict/path.png :) |
| 22:42 |
|
pdurbin |
(from http://martinfowler.com/bliki/MonolithFirst.html ) |
| 22:43 |
|
prologic |
omg |
| 22:43 |
|
prologic |
that's a bullshit diagram |
| 22:43 |
|
prologic |
who came up with this crap? |
| 22:44 |
|
prologic |
this is so against the UNIX Philosophy |
| 22:44 |
|
prologic |
geez |
| 22:44 |
|
prologic |
I know I have a brilliant idea |
| 22:44 |
|
prologic |
let's build the microwave oven with all the fancy programmable features with microships forst |
| 22:45 |
|
prologic |
before the simplest one |
| 22:45 |
|
prologic |
in almost any business; if you write the Monolithic program first |
| 22:45 |
|
prologic |
you'll be stuck with it forever and a day |
| 22:45 |
|
prologic |
it will be non-portable, unmaintainable |
| 22:46 |
|
prologic |
and businesses will not see value in rewriting it |
| 22:46 |
|
pdurbin |
well, he explains the diagram in the post I linked |
| 22:46 |
|
prologic |
small is beautiful |
| 22:46 |
|
prologic |
small is maintainable |
| 22:46 |
|
prologic |
I'll have a read |
| 22:46 |
|
prologic |
but so far I'm not impressed |
| 22:47 |
|
prologic |
"Monolithc first" is an outrageous thing to say |
| 22:47 |
|
pdurbin |
heh |
| 22:47 |
|
pdurbin |
well, at least he published the opposite opinion (differnt author) as well: http://martinfowler.com/articles/dont-start-monolith.html |
| 22:48 |
|
prologic |
hmm |
| 22:48 |
|
prologic |
so from that diagram he is advocating an iterative appproach to micirservices |
| 22:48 |
|
prologic |
such that you build up a big complex system first |
| 22:48 |
|
prologic |
and work out all the boundarieds |
| 22:48 |
|
prologic |
then split things up |
| 22:48 |
|
prologic |
I don't buy it |
| 22:49 |
|
prologic |
because like I said; you then have to rip things apart |
| 22:49 |
|
prologic |
businesses won't buy into it either and see no value in splitting it up |
| 22:49 |
|
prologic |
I'm not making this up |
| 22:49 |
|
prologic |
hard experience |
| 22:49 |
|
prologic |
:) |
| 22:49 |
|
prologic |
have enough foresight and design upfront to know what the all the pieces are up-front |
| 22:49 |
|
prologic |
build those and make them fit together |
| 22:49 |
|
prologic |
if you don't get it right the first time; "that's okay" |
| 22:50 |
|
prologic |
they're just components and can be swapped out easily if you write well defined programs with well defined interfaces |
| 22:53 |
|
pdurbin |
prologic: but perhaps developers more junior than you wouldn't work out all the boundaries. they'd be more comfortable building a monolith to start |
| 22:54 |
|
prologic |
hmm |
| 22:54 |
|
prologic |
perhaps you have a point there |
| 22:54 |
|
prologic |
but that's no excuse to be writing "bad software" |
| 22:54 |
|
prologic |
I think Juniors should be trained in the ways of the UNIX way |
| 22:54 |
|
prologic |
as such |
| 22:54 |
|
prologic |
we do actually teach this stuff in Uni btw |
| 22:54 |
|
prologic |
so I'm not sure why even Juniors wouldn't know better |
| 23:01 |
|
pdurbin |
"even experienced architects working in familiar domains have great difficulty getting boundaries right at the beginning" |
| 23:01 |
|
pdurbin |
anyway |
| 23:02 |
|
pdurbin |
I'd love to get better and building with microservices. I'm definitely a big unix fan. |
| 23:38 |
|
sivoais |
if you're writing unit tests, aren't you feeling out the boundaries anyway? |
| 23:40 |
|
sivoais |
oh, btw, re: my question about Dr. Dobb's Journal, InfoQ is a nice equivalent |
| 23:40 |
|
sivoais |
<http://www.infoq.com/> |
| 23:43 |
|
pdurbin |
yeah, I've seen some nice stuff on InfoQ |
| 23:43 |
|
pdurbin |
to be clear, I'm not arguing for monolith first :) |
| 23:43 |
|
pdurbin |
I just think there are some valid points in that article. |
