scala-virtualized, lms & stanford-ppl (delite, optiml, optiql)

[Daniel Mahler]

Has anybody looked at or even worked with scala-virtualized, lightweight modular staging and/or any of the stanford ppl libraries (delite, optiml, optiql, ...)?
We could add these to the list of potential meetup topics.

cheers
D

[Sukant Hajra]

Hi Daniel,

Sorry for the late reply. I saw the last batch of talk between you and Cody
about meetup topics and ran into an unfortunate hitch with how I was hoping to
manage this.

I'd originally hoped to have a permanently open poll on Meetup.com [1] but
Meetup limits the number of items you can have on the poll to only 10.

I could just make another poll with more topics -- a little janky, though.

I kind of like keeping people on the Meetup site, though, because the
federation seems convenient for people.

Any thoughts on how to better manage this?

Oh yeah. . . Cody, Daniel, I could really use a date from you guys about when
you want to present this month. If you don't reply back here, I'll email you
both in some private correspondence.

-Sukant

[1] http://www.meetup.com/Austin-Scala-Enthusiasts/polls/920802/

[Cody Koeninger]

I'm pretty flexible on dates.  I would prefer to avoid monday or wednesday, but can do that if necessary.

As far as the poll goes, can you remove items once we've done a meetup on that topic?

--
You received this message because you are subscribed to the Google Groups "Austin Scala Enthusiasts" group.
To unsubscribe from this group and stop receiving emails from it, send an email to austin-scala-enthu...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Sukant Hajra]

On 2013-11-06, Cody Koeninger wrote:
>
> I'm pretty flexible on dates. I would prefer to avoid monday or wednesday,
> but can do that if necessary.

I also don't prefer Monday or Wednesday. How about Thursday? I believe Austin
may have mentioned that he preferred Tuesday.

Is next week too soon to put something together (say Thursday Nov. 21st)?

Holidays are rough for both November and December, but I'd like to fit things
in if we can.

I've been a bit consumed with the transition to my new role, but things are
settling down, so I'll be more responsive to this now.

> As far as the poll goes, can you remove items once we've done a meetup on
> that topic?

Even after removing ones we've done (we've only met three times) it's easy to
fill up only 10 slots, especially with the new topics you and Daniel mentioned.

-Sukant

[Cody Koeninger]

Next thursday works for me.

[Sukant Hajra]

On 2013-11-12, Cody Koeninger wrote:
>
> Next thursday works for me.

Have you talked to Daniel? Is he okay with the date too? Are you guys still
slated to co-present?

-Sukant

[Cody Koeninger]

Haven't heard from him, apart from an initial email exchange the day after the prior meetup

-Sukant

[Daniel Mahler]

Hi Cody & Skant,

My time has been a lttle crazy lately.

Thurs looks good atm, but things are a little hard to predict.

We are talking about Nov 21 yes.

cheers

D

[Sukant Hajra]

On 2013-11-12, Daniel Mahler wrote:
>
> Thurs looks good atm, but things are a little hard to predict. We are
> talking about Nov 21 yes.

Yeah, next Thursday Nov. 21. If you two can coordinate about who will talk
about what first, then I can set up everything except for the pizza (which Cody
has generously offered to cover).

-Sukant

[Cody Koeninger]

I was going to talk about writing your own RDD, and other ways of doing ingress, which is probably better suited to going second

[Sukant Hajra]

On 2013-11-12, Cody Koeninger wrote:
>

> I was going to talk about writing your own RDD, and other ways of doing
> ingress, which is probably better suited to going second

Very interesting. Curious what your motivation for using a custom RDD is.
I've been investigating using the RDD for Hadoop input formats, and then using
the input format for Cassandra. This way, I hoping to have some benefits of a
real-time system from Cassandra, and benefits of a bulk-processing system from
Spark.

Also, I'm not familiar with the term "ingress."

So about the date, as soon as I get confirmation from Daniel that it's a good
date, I'll post the meetup. Or. . . Cody, if you think you have enough
material to do a full session (I suspect you do) I can safely just reserve the
room and send out notices now.

-Sukant

[Cody Koeninger]

By ingress i just mean input, ingest, get yer data in there.

When I first started working with spark, there wasn't an obvious way to get data from a relational database into spark, so I wrote an RDD for it, and it got accepted as a patch into spark.

My main point there being RDD is sufficiently general enough interface that you can get whatever data you need into spark, it's definitely not dependent on hadoop.  I'd like to go over the interface because I think it's a good jumping on point for the spark code.

We are using the newHadoopRdd as is to talk to cassandra, but there were a couple of interesting quirks to work out.  I know evan chan @ ooyala just ended up writing his own rdd for it - http://www.youtube.com/watch?v=6kHlArorzvs

I dunno if that's enough for a full session.  If people want a general overview, the ampcamp stuff is honestly better than whatever I'd do.

-Sukant

[Sukant Hajra]

On 2013-11-12, Cody Koeninger wrote:
>
> We are using the newHadoopRdd as is to talk to cassandra, but there were a
> couple of interesting quirks to work out.

Any chance I could bother you to learn about these interesting quirks? I'm
about to start prototyping stuff you already seem to have worked out, and I
think you might save me some time.

Feel free to tell more here. Or maybe I'll just talk to you after the meetup?

> I know evan chan @ ooyala just ended up writing his own rdd for it -
> http://www.youtube.com/watch?v=6kHlArorzvs

I watched that video recently, but wasn't sure how much of it was custom and
how much extension of APIs available from Cassandra/Spark/Hadoop.

Perhaps I should just ping Evan directly.

> I dunno if that's enough for a full session. If people want a general
> overview, the ampcamp stuff is honestly better than whatever I'd do.

Are you okay with me just setting up the meeting for Thursday next week, even
if Daniel isn't 100% prepared? Time slips so fast, and I'm worried about
giving everyone sufficient notice. If not, we'll just roll with it.

Thanks,
Sukant

[Cody Koeninger]

Yeah next thursday is fine, I can make sure to have some intro stuff about spark as well.

We can talk more about the cassandra stuff during/after the meetup, or if its time sensitive just email me.

On Tuesday, November 12, 2013, Sukant Hajra <2q2wd...@snkmail.com> wrote:

On 2013-11-12, Cody Koeninger wrote:
>
> We are using the newHadoopRdd as is to talk to cassandra, but there were a
> couple of interesting quirks to work out.

Any chance I could bother you to learn about these interesting quirks?  I'm
about to start prototyping stuff you already seem to have worked out, and I
think you might save me some time.

Feel free to tell more here.  Or maybe I'll just talk to you after the meetup?

> I know evan chan @ ooyala just ended up writing his own rdd for it -

> C* Summit 2013: Real-time Analytics using Cassandra, Spark and Shark

I watched that video recently, but wasn't sure how much of it was custom and
how much extension of APIs available from Cassandra/Spark/Hadoop.

Perhaps I should just ping Evan directly.

> I dunno if that's enough for a full session.  If people want a general
> overview, the ampcamp stuff is honestly better than whatever I'd do.

Are you okay with me just setting up the meeting for Thursday next week, even
if Daniel isn't 100% prepared?  Time slips so fast, and I'm worried about
giving everyone sufficient notice.  If not, we'll just roll with it.

Thanks,
Sukant

[Daniel Mahler]

Nearly there. Bad traffic.