| Time |
S |
Nick |
Message |
| 04:36 |
|
|
mr_lou joined ##friendlyjava |
| 14:03 |
|
pulsar |
aditsu: jackson and gson are the most popular ones IMHO. i have been using jackson for most of the time in the past |
| 14:04 |
|
pulsar |
in new projects without blueprint restrictions i have had no issues with gson and found the api a bit more intuitive |
| 14:04 |
|
pulsar |
but this is just an biased gut feeling. both frameworks are perfectly fine. |
| 14:30 |
|
aditsu |
ok, thanks |
| 14:32 |
|
pdurbin |
Right now we have JsonParser and JsonPrinter classes but it would be nice to move to JSR 367 once it's final. |
| 14:36 |
|
aditsu |
I used to use json-java (at least that's what it's called now, I think), but it's quite ugly |
| 14:38 |
|
pdurbin |
the thing is, we have business logic in those classes |
| 15:28 |
|
|
Guest40 joined ##friendlyjava |
| 15:49 |
|
aditsu |
I'm not dealing with classes in this case |
| 15:51 |
|
aditsu |
well, maybe just a simple class, but it only corresponds to a small part of the json data |
| 18:17 |
|
|
Guest40 joined ##friendlyjava |
| 19:30 |
|
pulsar |
aditsu: jaxb+jackson+@JsonIgnore might be something you want to look at |
| 19:32 |
|
aditsu |
xml binding? I'm pretty sure that's something I really don't want to look at |
| 19:33 |
|
pulsar |
it works well with json output/input too |
| 19:35 |
|
pulsar |
i mean, if you are dealing with a servlet container and want to consume/serve json messages - i would not do that by hand in that case |
| 19:35 |
|
pulsar |
otherwise it might not be the smartest choice, yes. |
| 22:32 |
|
pulsar |
i have been doing some work with spark the past week. ran a couple of jobs on a 40 node cluster. that and apache zeppelin: <3 |
| 22:33 |
|
pulsar |
first time i could use scala where it made actually some sense too. refreshing experience |
| 23:00 |
|
pdurbin |
pulsar: cool. People at work use Scala and Spark |
| 23:01 |
|
pulsar |
i have been using map reduce / hadoop / hbase for a couple of years now |
| 23:02 |
|
pulsar |
what a notebook like zeppelin does to your turnaround cycles and what scala+spark does to the lines of code when compared to classic mr jobs is just insane |
| 23:02 |
|
pdurbin |
they just gave a talk about it: https://github.com/ekraffmiller/JavaOne2016 |
| 23:03 |
|
pulsar |
uh, a bit sloppy formatting, eh? |
| 23:03 |
|
pulsar |
https://github.com/ekraffmiller/JavaOne2016/blob/master/src/main/java/edu/harvard/iq/javaone2016/AnalyzeTextFiles.java |
| 23:03 |
|
pulsar |
:) |
| 23:04 |
|
pulsar |
i am also amost sure, that java makes little sense if you do not have to deal with legacy libs / code when implementing spark jobs |
| 23:05 |
|
pdurbin |
they gave a practice talk. it was cool to see them scale up the work being done |
