Can extra processing threads help in this case?
in [MFC]
|
From: Hector Santos on 1 Apr 2010 15:10 Peter Olcott wrote: >> **** >> This is a possible approach, as long as you can guarantee >> that this cannot fail. >> joe >> *** > > That mostly depends upon the quality of my IPC design. I > posted this in the Unix/Linux/Threads groups because this > design depends on features of the Unix/Linux OS. > > Basically it has two named pipes that form the FIFO Queue, > and both of these coordinate through a single binary file > with fixed length records, that forms the transaction log. Ok, you can't do that on windows. Thats Linus technology. > It requires two named pipes one for each direction of > communication. The web server uses one pipe provide > transactions to the OCR process. The OCR process uses the > other pipe to inform the web server that one transaction has > been processed. Windows doesn't support Named Pipes! > As soon as a web request is made it is written to the log > file, and then to the FIFO Queue connected to the OCR > process. The OCR process then updates the transaction log > status to [Being Processed]. > > The log file keeps track of all of the transaction details, > and has three flags: > 0=Available for Processing > 1=Being Processed > 2=Processing is Completed You been reading my mail! > As soon as a web request has completed processing, the > transaction log status is updated to [Processing is > Completed], and a message is written to the FIFO Queue > connected to the web server. > > Unix/Linux guarantees that appends to files is an atomic > operation, and provides pread() and pwrite() that perform a > seek and then a read or write as a single atomic operation. > The log file is the persistent storage and serves many > purposes. Told you that too, but it doesn't work under Windows but its coming in Windows 2020 just in time for our vapor ware product! So why are you here? -- HLS
From: Peter Olcott on 1 Apr 2010 15:16 "Liviu" <lab2k1(a)gmail.c0m> wrote in message news:uFES1oc0KHA.3412(a)TK2MSFTNGP05.phx.gbl... > "Joseph M. Newcomer" <newcomer(a)flounder.com> wrote... >> On Wed, 31 Mar 2010 22:56:59 -0500, "Liviu" >> <lab2k1(a)gmail.c0m> wrote: >>>"Peter Olcott" <NoSpam(a)OCR4Screen.com> wrote... >>>> >>>> If you read much of the patent you would see that this >>>> is not the >>>> case. >>> >>>Oh, right... There is also the "store a DFA as a sparse >>>matrix" thing. >> **** >> Back in 1982, I implemented a new kind of sparse matrix >> approach, >> "comb vectors", based on a research paper from the >> University of >> Karlsruhe, Karlsruhe, Germany, which outlined the >> algorithm. I'm not >> sure if Gerhod Goos was one of the authors or it was a >> set of his >> students; I no longer have the paper. But it >> "interlaces" the "holes" >> in a parse-table vector so that we got incredible density >> in the >> immense parse tables. I still have the code, but cannot >> release it. >> So recompaction of spare matrices was an idea that was >> well-understood >> in 1981 or so. > > Sure was. I (used to) know a thing or two about sparse > matrices back in > the early 80's, having done some research which involved > resultants of > sparse polynomials. It was also obvious at the time that > the more one > knew about particular "sparseness" characteristics of the > matrix, the > more dramatic "compaction" could be achieved. > > Unrelated (and this is no stab at P.O.) but the patent > language in > general amuses me to no end with "de rigueur" convolutions > like... > > || The term "Sparse Matrix" is taken to have the common > meaning > || of the term "Sparse" combined with the common computer > science > || meaning of the term "Matrix", a two dimensional array > of elements. http://en.wikipedia.org/wiki/Sparse_matrix To be unambiguously distinguished from: In the subfield of numerical analysis, a sparse matrix is a matrix populated primarily with zeros (Stoer & Bulirsch 2002, p. 619). The term itself was coined by Harry M. Markowitz. > >> So it opens the issue about prior art on some of the >> compaction; >> I have not read the patent details soo I don't know if >> the compaction >> model differs from what I know to be past history, but I >> do know >> that DFA compaction is still patentable > > I never disputed that point. As for his patent, on cursory > reading it's > a combination between a fast DFA "exact pixel matching" > recognition > (compare to: face tagging in photos, 100% accurate if one > has a prior > database of the face at all sizes, angles, expressions, > lighting etc) > and an idea that it could be somehow used for "remote > control" of > another system (compare to: character-mode screen scraping > and > keyboard control over a terminal session into a > mainframe). I did not > mean to, and not going to, argue the merits, or the prior > art doubts. > The point I tried to make a few times (too many) was that > without > a viable implementation and/or business plan, the best of > patents will > forever remain just a piece of paper to frame and hang on > the wall. > > Liviu > > I have the design of an implementation that has very substantial benefits over any other comparable technology. These substantial benefits are very easy to map to significant dollar cost savings. My closest competitors are reportedly making millions, this report might be erroneous. Here is one of these close competitors: http://www.testplant.com/ The design has such substantial benefits over the close competitors that it would be able to compete against other (not so close) technologies as well.
From: Pete Delgado on 2 Apr 2010 10:57 "Hector Santos" <sant9442(a)nospam.gmail.com> wrote in message news:eIYKO8c0KHA.4412(a)TK2MSFTNGP02.phx.gbl... >> >> Basically it has two named pipes that form the FIFO Queue, and both of >> these coordinate through a single binary file with fixed length records, >> that forms the transaction log. > > > Ok, you can't do that on windows. Thats Linus technology. > >> It requires two named pipes one for each direction of communication. The >> web server uses one pipe provide transactions to the OCR process. The OCR >> process uses the other pipe to inform the web server that one transaction >> has been processed. > > > Windows doesn't support Named Pipes! > ???? Oh! I just noticed the posting date!!!! -Pete
From: Joseph M. Newcomer on 2 Apr 2010 11:31 See below... On Thu, 1 Apr 2010 14:16:30 -0500, "Peter Olcott" <NoSpam(a)OCR4Screen.com> wrote: > >"Liviu" <lab2k1(a)gmail.c0m> wrote in message >news:uFES1oc0KHA.3412(a)TK2MSFTNGP05.phx.gbl... >> "Joseph M. Newcomer" <newcomer(a)flounder.com> wrote... >>> On Wed, 31 Mar 2010 22:56:59 -0500, "Liviu" >>> <lab2k1(a)gmail.c0m> wrote: >>>>"Peter Olcott" <NoSpam(a)OCR4Screen.com> wrote... >>>>> >>>>> If you read much of the patent you would see that this >>>>> is not the >>>>> case. >>>> >>>>Oh, right... There is also the "store a DFA as a sparse >>>>matrix" thing. >>> **** >>> Back in 1982, I implemented a new kind of sparse matrix >>> approach, >>> "comb vectors", based on a research paper from the >>> University of >>> Karlsruhe, Karlsruhe, Germany, which outlined the >>> algorithm. I'm not >>> sure if Gerhod Goos was one of the authors or it was a >>> set of his >>> students; I no longer have the paper. But it >>> "interlaces" the "holes" >>> in a parse-table vector so that we got incredible density >>> in the >>> immense parse tables. I still have the code, but cannot >>> release it. >>> So recompaction of spare matrices was an idea that was >>> well-understood >>> in 1981 or so. >> >> Sure was. I (used to) know a thing or two about sparse >> matrices back in >> the early 80's, having done some research which involved >> resultants of >> sparse polynomials. It was also obvious at the time that >> the more one >> knew about particular "sparseness" characteristics of the >> matrix, the >> more dramatic "compaction" could be achieved. >> >> Unrelated (and this is no stab at P.O.) but the patent >> language in >> general amuses me to no end with "de rigueur" convolutions >> like... >> >> || The term "Sparse Matrix" is taken to have the common >> meaning >> || of the term "Sparse" combined with the common computer >> science >> || meaning of the term "Matrix", a two dimensional array >> of elements. > > http://en.wikipedia.org/wiki/Sparse_matrix >To be unambiguously distinguished from: > >In the subfield of numerical analysis, a sparse matrix is a >matrix populated primarily with zeros (Stoer & Bulirsch >2002, p. 619). The term itself was coined by Harry M. >Markowitz. **** Actually, a sparse (2-D, to simplify the discussion) matrix is a matrix of dimension [m*n] whose values at some coordinates [i,j] is not defined. 0.0 is a defined value, and therefore a matrix with 0 values is not "sparse". Essentially, a sparse matrix is a matrix for which the mapping [i x j] -> value is not a total mapping. A numerical analyst is free to redefine this in some other terms, but the essence of a sparse matrix is the failure to have a total mapping. The trick on "comb vectors" is to create an alternate representation that maintains the partial mapping of the original while implementing it as a vector such that the mapping [i]->value is nearly total (the difference represents the ability to do perfect compaction). So in the abstract world, a sparse 2-D matrix is characterized by a partial mapping (forall)i (forall)j [i X j]. I could try to write out the fully general characterization, but without using Microsoft Equation Editor it is clumsy. To interpret this any other way is to misinterpret the definition, or, as the numerical analysts have apparently redefined it, to change the definition to something else. joe > >> >>> So it opens the issue about prior art on some of the >>> compaction; >>> I have not read the patent details soo I don't know if >>> the compaction >>> model differs from what I know to be past history, but I >>> do know >>> that DFA compaction is still patentable >> >> I never disputed that point. As for his patent, on cursory >> reading it's >> a combination between a fast DFA "exact pixel matching" >> recognition >> (compare to: face tagging in photos, 100% accurate if one >> has a prior >> database of the face at all sizes, angles, expressions, >> lighting etc) >> and an idea that it could be somehow used for "remote >> control" of >> another system (compare to: character-mode screen scraping >> and >> keyboard control over a terminal session into a >> mainframe). I did not >> mean to, and not going to, argue the merits, or the prior >> art doubts. >> The point I tried to make a few times (too many) was that >> without >> a viable implementation and/or business plan, the best of >> patents will >> forever remain just a piece of paper to frame and hang on >> the wall. >> >> Liviu >> >> > >I have the design of an implementation that has very >substantial benefits over any other comparable technology. >These substantial benefits are very easy to map to >significant dollar cost savings. My closest competitors are >reportedly making millions, this report might be erroneous. > >Here is one of these close competitors: > http://www.testplant.com/ > >The design has such substantial benefits over the close >competitors that it would be able to compete against other >(not so close) technologies as well. > Joseph M. Newcomer [MVP] email: newcomer(a)flounder.com Web: http://www.flounder.com MVP Tips: http://www.flounder.com/mvp_tips.htm
From: Peter Olcott on 2 Apr 2010 11:33
"Pete Delgado" <Peter.Delgado(a)NoSpam.com> wrote in message news:%23iIPbTn0KHA.4420(a)TK2MSFTNGP02.phx.gbl... > > "Hector Santos" <sant9442(a)nospam.gmail.com> wrote in > message news:eIYKO8c0KHA.4412(a)TK2MSFTNGP02.phx.gbl... >>> >>> Basically it has two named pipes that form the FIFO >>> Queue, and both of these coordinate through a single >>> binary file with fixed length records, that forms the >>> transaction log. >> >> >> Ok, you can't do that on windows. Thats Linus >> technology. >> >>> It requires two named pipes one for each direction of >>> communication. The web server uses one pipe provide >>> transactions to the OCR process. The OCR process uses >>> the other pipe to inform the web server that one >>> transaction has been processed. >> >> >> Windows doesn't support Named Pipes! Whoops? http://msdn.microsoft.com/en-us/library/aa365590(VS.85).aspx >> > > ???? > > Oh! I just noticed the posting date!!!! > > -Pete > |