October 2008 - mpiwg-coll - mailman3.cels.anl.gov

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Rajeev Thakur 15 Oct '08

15 Oct '08

Or if we just drop nonblocking collectives altogether, the issue will not arise :-). Just kidding!! Rajeev > -----Original Message----- > From: mpi3-coll-bounces(a)lists.mpi-forum.org > [mailto:[email protected]] On Behalf Of > Darius Buntinas > Sent: Wednesday, October 15, 2008 8:59 PM > To: MPI-3 Collective Subgroup Discussions > Subject: Re: [mpi3-coll] Non-blocking Collectives Proposal Draft > > Torsten, perhaps we should follow the German style and call > it "nonblockingcollectives" :-) > > -d > > On 10/15/2008 08:27 PM, Torsten Hoefler wrote: > > > So as weird as it looks, I'd propose "nonblocking" as the > way to go (I > > guess I should like it as a German). Any opinions on this? > > > > Best, > > Torsten > > > _______________________________________________ > mpi3-coll mailing list > mpi3-coll(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll >

1 0

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Darius Buntinas 15 Oct '08

15 Oct '08

Torsten, perhaps we should follow the German style and call it "nonblockingcollectives" :-) -d On 10/15/2008 08:27 PM, Torsten Hoefler wrote: > So as weird as it looks, I'd propose "nonblocking" as the way to go (I > guess I should like it as a German). Any opinions on this? > > Best, > Torsten >

1 0

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Torsten Hoefler 15 Oct '08

15 Oct '08

On Wed, Oct 15, 2008 at 04:42:45PM -0700, Bronis R. de Supinski wrote: > > Rajeev and Torsten: > > I believe that Rajeev is correct. The Chicago Manual of > Style passge that Torsten cited is not germane. It deals > with whether you are supposed to add hyphens to adjective/ > noun combinations that are being used as adjectives, e.g., > "graduate student" (instead of "graduate-student") in > graduate student housing. It has nothing to do with > hyphenating compound words or words that are synthesized > by adding well defined prefixes to existing words. ok, I capitulate with the hyphen :). But after reading CMS 7.90 and 8.170, http://amaediting.blogspot.com/2008/07/hyphenated-prefixes.html and http://www.nyu.edu/classes/copyXediting/Hyphens.html , and talking to more native speakers, I would say that everything seems to be evolving to closed prefixes. So as weird as it looks, I'd propose "nonblocking" as the way to go (I guess I should like it as a German). Any opinions on this? Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- "Our two greatest problems are gravity and paper work. We can lick gravity, but sometimes the paperwork is overwhelming." -- Wernher von Braun

1 0

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Bronis R. de Supinski 15 Oct '08

15 Oct '08

Rajeev and Torsten: I believe that Rajeev is correct. The Chicago Manual of Style passge that Torsten cited is not germane. It deals with whether you are supposed to add hyphens to adjective/ noun combinations that are being used as adjectives, e.g., "graduate student" (instead of "graduate-student") in graduate student housing. It has nothing to do with hyphenating compound words or words that are synthesized by adding well defined prefixes to existing words. Bronis On Wed, 15 Oct 2008, Rajeev Thakur wrote: > I will just add that if you go to www. dictionary.com and look up > "nonbelligerent", it doesn't have a hyphen. In the left column, under > "Browse nearby entries" you can scroll through a boat load of "non" words > that don't have a hyphen. If you look up > http:// www. merriam-webster.com/dictionary/nonbelligerent, at the bottom it > shows the MLA style and APA style also without the hyphen. > > Rajeev > > > -----Original Message----- > > From: mpi3-coll-bounces(a)lists.mpi-forum.org > > [mailto:[email protected]] On Behalf Of > > Torsten Hoefler > > Sent: Wednesday, October 15, 2008 3:53 PM > > To: MPI-3 Collective Subgroup Discussions > > Subject: Re: [mpi3-coll] Non-blocking Collectives Proposal Draft > > > > On Wed, Oct 15, 2008 at 03:46:39PM -0500, Rajeev Thakur wrote: > > > We can do whatever we want really. It's not terribly > > important whether > > > nonblocking is hyphenated or not. > > that's right :). I will still add this as an MPI-2.2 ticket so that we > > can have the official 5 minute discussion in front of the forum. > > > > Thanks for your remarks! > > > > Best, > > Torsten > > > > -- > > bash$ :(){ :|:&};: --------------------- http:// www. unixer.de/ ----- > > Torsten Hoefler | Postdoctoral Researcher > > Open Systems Lab | Indiana University > > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > > Lindley Hall Room 135 | +01 (812) 855-3608 > > _______________________________________________ > > mpi3-coll mailing list > > mpi3-coll(a)lists.mpi-forum.org > > http:// lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll > > > > _______________________________________________ > mpi3-coll mailing list > mpi3-coll(a)lists.mpi-forum.org > http:// lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll > >

1 0

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Torsten Hoefler 15 Oct '08

15 Oct '08

On Wed, Oct 15, 2008 at 04:03:48PM -0500, Rajeev Thakur wrote: > I will just add that if you go to www.dictionary.com and look up > "nonbelligerent", it doesn't have a hyphen. In the left column, under > "Browse nearby entries" you can scroll through a boat load of "non" words > that don't have a hyphen. If you look up > http://www.merriam-webster.com/dictionary/nonbelligerent, at the bottom it > shows the MLA style and APA style also without the hyphen. yes, but this whole thing is very ambiguous. I entered non-blocking at www.dictionary.com which was not found, but the "Reference suggestions" show all three ways of writing it (1 x nonblocking, 1 x non blocking and 3 x non-blocking). And try to click on "non blocking" - it leads to a webpage with that only lists non-blocking. This is very unclear to me. Do you know a way to access MLA or APA online? This is getting interesting! I don't even know a rule how to do this in German. It would be interesting to know if there is a rule in English. Actually, try "cat /usr/share/dict/words | grep blocking" on your machine (not that this would mean anything more). I guess we will have (an amusing) vote on this in Chicago! All the Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Torsten Hoefler | Postdoctoral Researcher Open Systems Lab | Indiana University 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Rajeev Thakur 15 Oct '08

15 Oct '08

I will just add that if you go to www.dictionary.com and look up "nonbelligerent", it doesn't have a hyphen. In the left column, under "Browse nearby entries" you can scroll through a boat load of "non" words that don't have a hyphen. If you look up http://www.merriam-webster.com/dictionary/nonbelligerent, at the bottom it shows the MLA style and APA style also without the hyphen. Rajeev > -----Original Message----- > From: mpi3-coll-bounces(a)lists.mpi-forum.org > [mailto:[email protected]] On Behalf Of > Torsten Hoefler > Sent: Wednesday, October 15, 2008 3:53 PM > To: MPI-3 Collective Subgroup Discussions > Subject: Re: [mpi3-coll] Non-blocking Collectives Proposal Draft > > On Wed, Oct 15, 2008 at 03:46:39PM -0500, Rajeev Thakur wrote: > > We can do whatever we want really. It's not terribly > important whether > > nonblocking is hyphenated or not. > that's right :). I will still add this as an MPI-2.2 ticket so that we > can have the official 5 minute discussion in front of the forum. > > Thanks for your remarks! > > Best, > Torsten > > -- > bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- > Torsten Hoefler | Postdoctoral Researcher > Open Systems Lab | Indiana University > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > Lindley Hall Room 135 | +01 (812) 855-3608 > _______________________________________________ > mpi3-coll mailing list > mpi3-coll(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll >

1 0

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Torsten Hoefler 15 Oct '08

15 Oct '08

On Wed, Oct 15, 2008 at 03:46:39PM -0500, Rajeev Thakur wrote: > We can do whatever we want really. It's not terribly important whether > nonblocking is hyphenated or not. that's right :). I will still add this as an MPI-2.2 ticket so that we can have the official 5 minute discussion in front of the forum. Thanks for your remarks! Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Torsten Hoefler | Postdoctoral Researcher Open Systems Lab | Indiana University 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Rajeev Thakur 15 Oct '08

15 Oct '08

We can do whatever we want really. It's not terribly important whether nonblocking is hyphenated or not. Rajeev > -----Original Message----- > From: mpi3-coll-bounces(a)lists.mpi-forum.org > [mailto:[email protected]] On Behalf Of > Torsten Hoefler > Sent: Wednesday, October 15, 2008 3:38 PM > To: MPI-3 Collective Subgroup Discussions > Subject: Re: [mpi3-coll] Non-blocking Collectives Proposal Draft > > On Wed, Oct 15, 2008 at 03:24:49PM -0500, Rajeev Thakur wrote: > > I just walked over to our editor, Gail Pieper, and asked > her the reason. She > > said that if the alphabets that the prefix connects are > different, as in > > nonblocking or noncontiguous, the convention nowadays is to > drop the hyphen. > > You would keep the hyphen in words like re-establish and > also in words that > > would read weird without the hyphen, like non-organic (nono > rganic). This is > > the convention followed by the Chicago Manual of Style, the > Associated Press > > Manual, and the MLA Manual (Modern Language Association). > hmm, I just read your cited Chicago Manual of Style (CMS) and found: > http://www.chicagomanualofstyle.org/ch07/ch07_sec085.html > which reads: > "A hyphen can make for easier reading by showing structure and, often, > pronunciation." ... "Where no ambiguity could result, as in public > welfare administration or graduate student housing, hyphenation is not > mandatory, though it is quite acceptable and preferred by many writers > and editors." > > (you should read the complete page) > > > > Also, if you do the <pick your favorit search enigne> > test (I know ... > > > but languages are dynamically changed by their users), > you'll see that > > > non-blocking seems to be the most-used term in the I/O and > > > communication > > > context. > > > > Web searches are not useful for this because most people's > English is not > > that good. > I strongly disagree. A language is determined by the people > who use it, > who defines what the correct language is? Even the CMS says > things like: > "With frequent use" or "have become widely accepted,". See > http://www.chicagomanualofstyle.org/ch07/ch07_sec084.html for more of > this :). I know, especially as a German, that a clearly > defined language > is preferred, but I think our case is not clearly defined (or can you > cite a particular part of CMS?). > > If we don't have a clear definition, I'd like to go with the > unofficial > conventions. Those unofficial conventions can be backed up with web > searches reasonably well (I read the first few results and those were > reasonable hits). > > All the Best, > Torsten > > -- > bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- > Torsten Hoefler | Postdoctoral Researcher > Open Systems Lab | Indiana University > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > Lindley Hall Room 135 | +01 (812) 855-3608 > _______________________________________________ > mpi3-coll mailing list > mpi3-coll(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll >

1 0

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Torsten Hoefler 15 Oct '08

15 Oct '08

On Wed, Oct 15, 2008 at 03:24:49PM -0500, Rajeev Thakur wrote: > I just walked over to our editor, Gail Pieper, and asked her the reason. She > said that if the alphabets that the prefix connects are different, as in > nonblocking or noncontiguous, the convention nowadays is to drop the hyphen. > You would keep the hyphen in words like re-establish and also in words that > would read weird without the hyphen, like non-organic (nono rganic). This is > the convention followed by the Chicago Manual of Style, the Associated Press > Manual, and the MLA Manual (Modern Language Association). hmm, I just read your cited Chicago Manual of Style (CMS) and found: http://www.chicagomanualofstyle.org/ch07/ch07_sec085.html which reads: "A hyphen can make for easier reading by showing structure and, often, pronunciation." ... "Where no ambiguity could result, as in public welfare administration or graduate student housing, hyphenation is not mandatory, though it is quite acceptable and preferred by many writers and editors." (you should read the complete page) > > Also, if you do the <pick your favorit search enigne> test (I know ... > > but languages are dynamically changed by their users), you'll see that > > non-blocking seems to be the most-used term in the I/O and > > communication > > context. > > Web searches are not useful for this because most people's English is not > that good. I strongly disagree. A language is determined by the people who use it, who defines what the correct language is? Even the CMS says things like: "With frequent use" or "have become widely accepted,". See http://www.chicagomanualofstyle.org/ch07/ch07_sec084.html for more of this :). I know, especially as a German, that a clearly defined language is preferred, but I think our case is not clearly defined (or can you cite a particular part of CMS?). If we don't have a clear definition, I'd like to go with the unofficial conventions. Those unofficial conventions can be backed up with web searches reasonably well (I read the first few results and those were reasonable hits). All the Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Torsten Hoefler | Postdoctoral Researcher Open Systems Lab | Indiana University 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0

Re: [mpi3-coll] Non-blocking Collectives Proposal Draft
by Rajeev Thakur 15 Oct '08

15 Oct '08

I just walked over to our editor, Gail Pieper, and asked her the reason. She said that if the alphabets that the prefix connects are different, as in nonblocking or noncontiguous, the convention nowadays is to drop the hyphen. You would keep the hyphen in words like re-establish and also in words that would read weird without the hyphen, like non-organic (nono rganic). This is the convention followed by the Chicago Manual of Style, the Associated Press Manual, and the MLA Manual (Modern Language Association). > Also, if you do the <pick your favorit search enigne> test (I know ... > but languages are dynamically changed by their users), you'll see that > non-blocking seems to be the most-used term in the I/O and > communication > context. Web searches are not useful for this because most people's English is not that good. Rajeev > -----Original Message----- > From: mpi3-coll-bounces(a)lists.mpi-forum.org > [mailto:[email protected]] On Behalf Of > Torsten Hoefler > Sent: Wednesday, October 15, 2008 3:00 PM > To: MPI-3 Collective Subgroup Discussions > Subject: Re: [mpi3-coll] Non-blocking Collectives Proposal Draft > > On Wed, Oct 15, 2008 at 02:49:03PM -0500, Rajeev Thakur wrote: > > > 1) The MPI-2.1 document has 127 occurrences of "nonblocking" > > > and only 8 > > > occurrences of "non-blocking". What would be the correct > > > term? Should we try to be more consistent? > > > > The technical editor in our division prefers the version > without the hyphen. > > She will remove the hyphen in our papers. > > hmm, I would disagree - I had exactly the same discussion with our > technical editor :). She explained me for 15 minutes why it is > "non-blocking" and not "nonblocking" or "non blocking". > > Also, if you do the <pick your favorit search enigne> test (I know ... > but languages are dynamically changed by their users), you'll see that > non-blocking seems to be the most-used term in the I/O and > communication > context. > > I'm open for arguments! > > Best, > Torsten > > -- > bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- > Torsten Hoefler | Postdoctoral Researcher > Open Systems Lab | Indiana University > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > Lindley Hall Room 135 | +01 (812) 855-3608 > _______________________________________________ > mpi3-coll mailing list > mpi3-coll(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll >

1 0