According to slogan of project, we can clearly discern the core and non-core of Hbase; A Google BigTable clone. Then, we can think all things of the hbase shell are exists on another layer.
I think REST can be show that if a work environment is AJAX or some kind of open APIs, productivity increases. The REST approches can't be adopted for all types of application. And, i think sparse linear algebra is core of BigTable. ------------------------------ B. Regards, Edward yoon @ NHN, corp. Home : http://www.udanax.org > From: [EMAIL PROTECTED] > To: hadoop-dev@lucene.apache.org > Date: Fri, 7 Dec 2007 22:12:24 -0800 > Subject: RE: [VOTE][PLAN] Graduate Hbase Shell as a contrib project > > Just my two cents... > > The HBase shell has some components that are extremely valuable to > the core of HBase. In particular, administrative functions such as > creating/deleting/modifying tables and the ability to do simple > queries from the command line. > > On the other hand, some of the proposed extensions are either > extraneous or outside the scope of the HBase core. > > For example a socket server to the shell seems extraneous given > that HBase has a REST interface. High volume updates or queries > might better be handled by providing native language bindings for > HBase. One mechanism that has been proposed for this is to use > Thrift to communicate between clients and the HBase servers. > > Other extensions such as Matrix operations, RDF, and Relation-Algebra > might better be packaged in a separate jar so they could be > utilized in map/reduce jobs. Additionally, if they were > generalized so that they could accept inputs from sources other > than HBase, such as SequenceFiles (for example) it would be > possible to apply these operations to non-HBase map/reduce > jobs. In the shell, the ability to create and submit a map/reduce > job using these operations might also be useful. > > --- > Jim Kellerman, Senior Engineer; Powerset > > >> -----Original Message----- >> From: edward yoon [mailto:[EMAIL PROTECTED] >> Sent: Friday, December 07, 2007 9:36 PM >> To: hadoop-dev@lucene.apache.org >> Subject: RE: [VOTE][PLAN] Graduate Hbase Shell as a contrib project >> >> >> Ok stack, The effort to questionnaire rebuild has just begun. >> >> ------------------------------ >> >> B. Regards, >> >> Edward yoon @ NHN, corp. >> Home : http://www.udanax.org >> >> >>> Date: Fri, 7 Dec 2007 21:24:42 -0800 >>> From: [EMAIL PROTECTED] >>> To: hadoop-dev@lucene.apache.org >>> Subject: Re: [VOTE][PLAN] Graduate Hbase Shell as a contrib project >>> >>> Hey Edward, how about we consult first before putting out a >> call for a >>> vote. At least one option is missing as far as I can see: A >> yes/no on >>> whether patches and wiki proposals for >>> Matrix+RDF+Relation-Algebra+HQL-Server proposed shell extensions >>> Matrix+RDF+Relation-Algebra+should >>> be committed to hbase core. Other folks might want to add >> other options. >>> >>> St.Ack >>> >>> >>> edward yoon wrote: >>>> Hi, community. >>>> I'm edward yoon, a worker for the 'hbase shell on hbase' >> and 'RDF, Matrix storage on hadoop + hbase'. >>>> I and other guys who did register for the hbase shell >> project would graduate as a contrib project from hbase sub-package. >>>> I believe that It will makes the hadoop an even better >> platform for scientific and advanced analytics programming. >>>> >>>> So far, i was work alone for try to become a committer. >>>> But, it seems to slow way. >>>> I'd like to contribute these at once after internal works >> after set up and move the hbase shell code to somewhere under >> contrib package. >>>> >>>> So, i propose to graduate hbase shell as a contrib project. >>>> If you find portraits of interest, please vote for the >> contrib project. >>>> >>>> 1. http://wiki.apache.org/lucene-hadoop/Hbase/HbaseShell >>>> 2. http://wiki.apache.org/lucene-hadoop/Hbase/RDF >>>> 3. http://wiki.apache.org/lucene-hadoop/Hbase/Matrix >>>> >>>> A significant majority voted against the motion, I'll just entrust >>>> the hbase shell policy to Hbase committers because the >> complaints of the hbase shell in the hbase is much enough on >> hadoop list and chatting room. >>>> >>>> 1. Risky because it will encourage people to think that >> the shell is >>>> a good way to interact with HBase in general... [Bryan Duxbury] 2. >>>> Shell has outgrown hbase and no longer fits. It belongs in its own >>>> project.... [Michael Stack] 3. Why not startup a new project named >>>> hbaseshell over on sourceforge?... [Michael Stack] >>>> >>>> here's mine : >>>> >> --------------------------------------------------------------------- >>>> ---------------- [ ] +1 Remove the hbase shell to other contrib >>>> project from hbase [ ] +0 [ ] -0 [ ] -1 Retain the hbase >> shell as a >>>> package on hbase >>>> >> --------------------------------------------------------------------- >>>> ---------------- >>>> >>>> Thanks, >>>> Edward. >>>> >>>> ------------------------------ >>>> B. Regards, >>>> >>>> Edward yoon @ NHN, corp. >>>> Home : http://www.udanax.org >>>> _________________________________________________________________ >>>> Put your friends on the big screen with Windows Vista(r) + >> Windows Live(tm). >>>> >> http://www.microsoft.com/windows/shop/specialoffers.mspx?ocid=TXT_TAG >>>> LM_CPC_MediaCtr_bigscreen_102007 >>> >> >> _________________________________________________________________ >> Put your friends on the big screen with Windows Vista(r) + >> Windows Live(tm). >> http://www.microsoft.com/windows/shop/specialoffers.mspx?ocid= > TXT_TAGLM_CPC_MediaCtr_bigscreen_102007 >> _________________________________________________________________ Put your friends on the big screen with Windows Vista® + Windows Live™. http://www.microsoft.com/windows/shop/specialoffers.mspx?ocid=TXT_TAGLM_CPC_MediaCtr_bigscreen_102007
