hdt-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Evert Lammerts <ev...@apache.org>
Subject Fwd: [VOTE] Retire HDT
Date Sun, 23 Nov 2014 19:49:01 GMT
A little update to prevent the vote from closing: we're still discussing
possible steps to make, see details in the thread below. We should have
some sort of plan in a couple of days. In the meantime: contributions are
very welcome!


---------- Forwarded message ---------
From: Evert Lammerts <evert.lammerts@gmail.com>
Date: Sun Nov 23 2014 at 2:05:27 PM
Subject: Re: [VOTE] Retire HDT
To: Bob Kerns <rwk@acm.org>, Mirko Kämpf <mirko.kaempf@gmail.com>

I think that an environment that allows fast experimentation adds most
value: building MapReduce and Spark jobs through Java, Scala, Python and
Pig. It can provide a number of commonly used libraries (Mahout, Giraph,
MLlib, GraphX), UDF's, Input- OutputFormats, Pig Readers and Writers and
code templates. Would you guys commit to making an attempt at building
this? If so then I propose we write a short proposal and send it to the
list. In the meantime I'll forward this thread.

Bob, I don't know enough about virtualenv internals. Do you think it'd be
possible to plug that into Orion or Cloud9's server?

Mirko, I'm open to exploring possibilities within Oozie, but personally
I've never used it: I just write drivers that chain jobs. So I can't judge
its applicability right now. Could we create a standalone IDE and when (if)
we get that to work, plug it into Oozie?

How about effort? I'm willing to commit to putting some time in the coming
few months, but it'll be hard to do more than 10 hours weekly or so.



  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message