Re: [9fans] TeX: hurrah!
Hi I'm aware of XeTeX (I had mentioned XeLaTeX in an earlier thread), and yes, I understand one wouldn't be looking for identity with what other platforms support. I agree that one shouldn't be looking to ape, but rather to provide the same or more functionality in a better way. Perhaps I'll be able to help with that in due course. K Joel C. Salomon joelcsalo...@gmail.com 18/04/2010 12:24 am There are a few projects (in the TeX world) for that, primarily XeTeX and Omega. Omega is not much in use anymore, but its ideas live on in LuaTeX. The lack of C++ is going to hinder efforts to port these projects to Plan 9 as-is; and these are significant efforts, not likely to be duplicated by 9fans. (Perhaps the C++ library for PDF handling can be rewritten in C, and then XeTeX LuaTeX can be ported. But don't expect the projects to use the rewrite in favor of the original libraries.) —Joel Salomon
Re: [9fans] TeX: hurrah!
On Fri, Apr 16, 2010 at 6:10 PM, Karljurgen Feuerherm kfeuerh...@wlu.ca wrote: It occurred to me that a profitable thing to do here would be to mention some things that would be nice to see in a new improved TeX... I believe bidirectional was mentioned already. The other thing that is essential for folk like me is complete Unicode compatibility [Yes, I know. UTC has committed many sins :] There are a few projects (in the TeX world) for that, primarily XeTeX and Omega. Omega is not much in use anymore, but its ideas live on in LuaTeX. The lack of C++ is going to hinder efforts to port these projects to Plan 9 as-is; and these are significant efforts, not likely to be duplicated by 9fans. (Perhaps the C++ library for PDF handling can be rewritten in C, and then XeTeX LuaTeX can be ported. But don't expect the projects to use the rewrite in favor of the original libraries.) —Joel Salomon
Re: [9fans] TeX: hurrah!
On Fri, Apr 16, 2010 at 01:57:56PM +0200, tlaro...@polynum.com wrote: ... Nice work. Can't wait to try it. -- I am a man who does not exist for others. pgp5W3AtIaNFY.pgp Description: PGP signature
Re: [9fans] TeX: hurrah!
Nice work! On Fri, Apr 16, 2010 at 4:57 AM, tlaro...@polynum.com wrote: So it compiles without ado under Plan9! And it's pure C89 (POSIX is just for the framework, not for the code: I have removed unneeded dependencies). And it's all the latest versions of the programs. So some numbers: - You will need to download a bundle of 4 chunks (I will put all on my website; I think the licence for TeX/METAFONT allows thie): 1) Donald E. Knuth's sources: all the programs; the sources for the Computer Modern Typefaces and the man pages (roff) I have found here and there or reconstructed. = 6 Mb (compressed) 2) kertex_M: these are the tools for the _matrix_ (the vulcan; the node compiling; the cpu) translating from Pascal to C. = 54 kb (compressed) 3) kertex_T: the R.I.S.K. (see below) framework for building TeX/METAFONT, that is the change files, the template Makefiles, the PASCAL-H extensions implemented as a C library. Plus some public domain software: dvips [it is not compiled for now; it's the next in the list] and some auxiliary tools (afm2tfm etc.); MetaPost (not included for now: next after dvips). = 0.4 Mb 4) The R.I.S.K. comp framework (my stuff) that is just a basic POSIX compliant way of configuring, compiling and installing software. Everything is configured via basic parameters file (you should not have to change something here; but it's all sh(1), ed(1), sed(1) and make(1)). = 125kb (compressed; the bigger par in size is D.E. Knuth and Silvio Levy's CWEB [not used here] that I use intensively). So the sources, it is less than 12 Mb uncompressed. And the sources can be (shall be) read-only so you can put them where you want. Space requirements for compilation : you will need about 33 Mb of free space or, at least, 12 Mb of free space if you use the SAVE_SPACE=YES option of R.I.S.K. (in this case, the intermediary products are removed whence the target is built; and when building the package, the binaries are not copied in the tarball, but moved in the tarball). Time: on my old AMD Athlon (686-class), 1250.16 MHz, this takes less than 2 minutes and a half... The resulting package is 8 Mb. But the installation will need a little more, since : 1) The METAFONT plain format is compiled leading to the program that has the right to be called mf(1) since it passes the TRAP test. 2) All the Computer Modern fontes are compiled, using mf(1). 3) The TeX plain format is compiled leading to the program that has the right to be called tex(1) since it passes the TRIP test. Since the work is done for allowing cross-compilation, the binaries are not used on the matrix, but on the target, i.e. at installation time. For the use of the stuff, I will need to put the essential in the document describing the process in the following days. For Plan9, one question: about the group (since for the user it will be ignored). Which group ? I have decided that the stuff will be installed in: - /lib/kertex for machine independant stuff (the macros, but the compiled fontes too). Plan9 users can bind -a to ``redirect'' what will be written where they want. - /$objtype/bin for the binaries. (idem) - /$objtype/lib/kertex/tex/dump/, and /$objtype/lib/kertex/mf/dump/ for the dump of the tables (the compiled version of the macros, that is loaded say when you call virtex(1) mistex(1), this is virtex(1) but it looks to is dump mistex.fmt to load it. The dumps are not machine independant, even not program independant since they depend on options selected at compile time). (idem) To give you a view of what is installed (and possibly where), here is the (generic: for Unices too) map [I need to change root:wheel to variable things) : #== THE MAIN DIR TREE # Everything is relative (for the not man nor bin stuff) to PKGDIR that # one can set as he sees fit. # + d * $PKGDIR/mf root:wheel 755 # This is where we put Computer Modern _sources_ (parameters and drivers # METAFONT's files). # + d * $PKGDIR/mf/cm root:wheel 755 + d * $PKGDIR/mf/mac root:wheel 755 # If you want to ``localize'' messages, look at D.E.K.'s pool stuff. # + d * $PKGDIR/mf/pool root:wheel 755 + d * $PKGDIR/mf/trap root:wheel 755 # Since to create a METAFONT' base file, one dumps, we call it: dump. # But is is machine dependant. # + d * $TARGETBINDIR/lib root:wheel 755 + d * $TARGETBINDIR/lib/mf root:wheel 755 + d * $TARGETBINDIR/lib/mf/dump root:wheel
Re: [9fans] TeX: hurrah!
Hello! Congratulations! Have you any plans to adapt the TeX for UTF-8 input? On Fri, 16 Apr 2010 15:57:56 +0400, tlaro...@polynum.com tlaro...@polynum.com wrote: So it compiles without ado under Plan9! And it's pure C89 (POSIX is just for the framework, not for the code: I have removed unneeded dependencies). And it's all the latest versions of the programs. So some numbers: - You will need to download a bundle of 4 chunks (I will put all on my website; I think the licence for TeX/METAFONT allows thie): 1) Donald E. Knuth's sources: all the programs; the sources for the Computer Modern Typefaces and the man pages (roff) I have found here and there or reconstructed. = 6 Mb (compressed) 2) kertex_M: these are the tools for the _matrix_ (the vulcan; the node compiling; the cpu) translating from Pascal to C. = 54 kb (compressed) 3) kertex_T: the R.I.S.K. (see below) framework for building TeX/METAFONT, that is the change files, the template Makefiles, the PASCAL-H extensions implemented as a C library. Plus some public domain software: dvips [it is not compiled for now; it's the next in the list] and some auxiliary tools (afm2tfm etc.); MetaPost (not included for now: next after dvips). = 0.4 Mb 4) The R.I.S.K. comp framework (my stuff) that is just a basic POSIX compliant way of configuring, compiling and installing software. Everything is configured via basic parameters file (you should not have to change something here; but it's all sh(1), ed(1), sed(1) and make(1)). = 125kb (compressed; the bigger par in size is D.E. Knuth and Silvio Levy's CWEB [not used here] that I use intensively). So the sources, it is less than 12 Mb uncompressed. And the sources can be (shall be) read-only so you can put them where you want. Space requirements for compilation : you will need about 33 Mb of free space or, at least, 12 Mb of free space if you use the SAVE_SPACE=YES option of R.I.S.K. (in this case, the intermediary products are removed whence the target is built; and when building the package, the binaries are not copied in the tarball, but moved in the tarball). Time: on my old AMD Athlon (686-class), 1250.16 MHz, this takes less than 2 minutes and a half... The resulting package is 8 Mb. But the installation will need a little more, since : 1) The METAFONT plain format is compiled leading to the program that has the right to be called mf(1) since it passes the TRAP test. 2) All the Computer Modern fontes are compiled, using mf(1). 3) The TeX plain format is compiled leading to the program that has the right to be called tex(1) since it passes the TRIP test. Since the work is done for allowing cross-compilation, the binaries are not used on the matrix, but on the target, i.e. at installation time. For the use of the stuff, I will need to put the essential in the document describing the process in the following days. For Plan9, one question: about the group (since for the user it will be ignored). Which group ? I have decided that the stuff will be installed in: - /lib/kertex for machine independant stuff (the macros, but the compiled fontes too). Plan9 users can bind -a to ``redirect'' what will be written where they want. - /$objtype/bin for the binaries. (idem) - /$objtype/lib/kertex/tex/dump/, and /$objtype/lib/kertex/mf/dump/ for the dump of the tables (the compiled version of the macros, that is loaded say when you call virtex(1) mistex(1), this is virtex(1) but it looks to is dump mistex.fmt to load it. The dumps are not machine independant, even not program independant since they depend on options selected at compile time). (idem) To give you a view of what is installed (and possibly where), here is the (generic: for Unices too) map [I need to change root:wheel to variable things) : #== THE MAIN DIR TREE # Everything is relative (for the not man nor bin stuff) to PKGDIR that # one can set as he sees fit. # + d * $PKGDIR/mf root:wheel 755 # This is where we put Computer Modern _sources_ (parameters and drivers # METAFONT's files). # + d * $PKGDIR/mf/cm root:wheel 755 + d * $PKGDIR/mf/mac root:wheel 755 # If you want to ``localize'' messages, look at D.E.K.'s pool stuff. # + d * $PKGDIR/mf/pool root:wheel 755 + d * $PKGDIR/mf/trap root:wheel 755 # Since to create a METAFONT' base file, one dumps, we call it: dump. # But is is machine dependant. # + d * $TARGETBINDIR/lib
Re: [9fans] TeX: hurrah!
LaTeX does too. I have used it with a recent version of TeX Live to get greek and mathematical symbols in verbatim code listings. James On Fri, Apr 16, 2010 at 5:07 PM, Karljurgen Feuerherm kfeuerh...@wlu.ca wrote: XeTeX/XeLaTeX do this, I believe... Perhaps they can be ported at some point? K Karljürgen G. Feuerherm, PhD Department of Archaeology and Classical Studies Wilfrid Laurier University 75 University Avenue West Waterloo, Ontario N2L 3C5 Tel. (519) 884-1970 x3193 Fax (519) 883-0991 (ATTN Arch. Classics) Alexander Sychev santu...@gmail.com 16/04/2010 9:32:04 am Hello! Congratulations! Have you any plans to adapt the TeX for UTF-8 input? On Fri, 16 Apr 2010 15:57:56 +0400, tlaro...@polynum.com tlaro...@polynum.com wrote: So it compiles without ado under Plan9! And it's pure C89 (POSIX is just for the framework, not for the code: I have removed unneeded dependencies). And it's all the latest versions of the programs. So some numbers: - You will need to download a bundle of 4 chunks (I will put all on my website; I think the licence for TeX/METAFONT allows thie): 1) Donald E. Knuth's sources: all the programs; the sources for the Computer Modern Typefaces and the man pages (roff) I have found here and there or reconstructed. = 6 Mb (compressed) 2) kertex_M: these are the tools for the _matrix_ (the vulcan; the node compiling; the cpu) translating from Pascal to C. = 54 kb (compressed) 3) kertex_T: the R.I.S.K. (see below) framework for building TeX/METAFONT, that is the change files, the template Makefiles, the PASCAL-H extensions implemented as a C library. Plus some public domain software: dvips [it is not compiled for now; it's the next in the list] and some auxiliary tools (afm2tfm etc.); MetaPost (not included for now: next after dvips). = 0.4 Mb 4) The R.I.S.K. comp framework (my stuff) that is just a basic POSIX compliant way of configuring, compiling and installing software. Everything is configured via basic parameters file (you should not have to change something here; but it's all sh(1), ed(1), sed(1) and make(1)). = 125kb (compressed; the bigger par in size is D.E. Knuth and Silvio Levy's CWEB [not used here] that I use intensively). So the sources, it is less than 12 Mb uncompressed. And the sources can be (shall be) read-only so you can put them where you want. Space requirements for compilation : you will need about 33 Mb of free space or, at least, 12 Mb of free space if you use the SAVE_SPACE=YES option of R.I.S.K. (in this case, the intermediary products are removed whence the target is built; and when building the package, the binaries are not copied in the tarball, but moved in the tarball). Time: on my old AMD Athlon (686-class), 1250.16 MHz, this takes less than 2 minutes and a half... The resulting package is 8 Mb. But the installation will need a little more, since : 1) The METAFONT plain format is compiled leading to the program that has the right to be called mf(1) since it passes the TRAP test. 2) All the Computer Modern fontes are compiled, using mf(1). 3) The TeX plain format is compiled leading to the program that has the right to be called tex(1) since it passes the TRIP test. Since the work is done for allowing cross-compilation, the binaries are not used on the matrix, but on the target, i.e. at installation time. For the use of the stuff, I will need to put the essential in the document describing the process in the following days. For Plan9, one question: about the group (since for the user it will be ignored). Which group ? I have decided that the stuff will be installed in: - /lib/kertex for machine independant stuff (the macros, but the compiled fontes too). Plan9 users can bind -a to ``redirect'' what will be written where they want. - /$objtype/bin for the binaries. (idem) - /$objtype/lib/kertex/tex/dump/, and /$objtype/lib/kertex/mf/dump/ for the dump of the tables (the compiled version of the macros, that is loaded say when you call virtex(1) mistex(1), this is virtex(1) but it looks to is dump mistex.fmt to load it. The dumps are not machine independant, even not program independant since they depend on options selected at compile time). (idem) To give you a view of what is installed (and possibly where), here is the (generic: for Unices too) map [I need to change root:wheel to variable things) : #== THE MAIN DIR TREE # Everything is relative (for the not man nor bin stuff) to PKGDIR that # one can set as he sees fit. # + d * $PKGDIR/mf root:wheel 755 # This is where we put Computer Modern _sources_ (parameters and drivers # METAFONT's files). # + d * $PKGDIR/mf/cm root:wheel 755 + d * $PKGDIR/mf/mac root:wheel 755 # If you want to ``localize'' messages, look at D.E.K.'s pool stuff. # + d * $PKGDIR/mf/pool root:wheel 755 + d * $PKGDIR/mf/trap root:wheel 755 # Since to create a METAFONT' base
Re: [9fans] TeX: hurrah!
On Fri, 16 Apr 2010 18:07:14 +0400, Karljurgen Feuerherm kfeuerh...@wlu.ca wrote: XeTeX/XeLaTeX do this, I believe... Perhaps they can be ported at some point? IFAIK, XeTeX/XeLaTeX based on C++ code. K Karljürgen G. Feuerherm, PhD Department of Archaeology and Classical Studies Wilfrid Laurier University 75 University Avenue West Waterloo, Ontario N2L 3C5 Tel. (519) 884-1970 x3193 Fax (519) 883-0991 (ATTN Arch. Classics) Alexander Sychev santu...@gmail.com 16/04/2010 9:32:04 am Hello! Congratulations! Have you any plans to adapt the TeX for UTF-8 input? On Fri, 16 Apr 2010 15:57:56 +0400, tlaro...@polynum.commailto:tlaro...@polynum.com tlaro...@polynum.commailto:tlaro...@polynum.com wrote: So it compiles without ado under Plan9! And it's pure C89 (POSIX is just for the framework, not for the code: I have removed unneeded dependencies). And it's all the latest versions of the programs. So some numbers: - You will need to download a bundle of 4 chunks (I will put all on my website; I think the licence for TeX/METAFONT allows thie): 1) Donald E. Knuth's sources: all the programs; the sources for the Computer Modern Typefaces and the man pages (roff) I have found here and there or reconstructed. = 6 Mb (compressed) 2) kertex_M: these are the tools for the _matrix_ (the vulcan; the node compiling; the cpu) translating from Pascal to C. = 54 kb (compressed) 3) kertex_T: the R.I.S.K. (see below) framework for building TeX/METAFONT, that is the change files, the template Makefiles, the PASCAL-H extensions implemented as a C library. Plus some public domain software: dvips [it is not compiled for now; it's the next in the list] and some auxiliary tools (afm2tfm etc.); MetaPost (not included for now: next after dvips). = 0.4 Mb 4) The R.I.S.K. comp framework (my stuff) that is just a basic POSIX compliant way of configuring, compiling and installing software. Everything is configured via basic parameters file (you should not have to change something here; but it's all sh(1), ed(1), sed(1) and make(1)). = 125kb (compressed; the bigger par in size is D.E. Knuth and Silvio Levy's CWEB [not used here] that I use intensively). So the sources, it is less than 12 Mb uncompressed. And the sources can be (shall be) read-only so you can put them where you want. Space requirements for compilation : you will need about 33 Mb of free space or, at least, 12 Mb of free space if you use the SAVE_SPACE=YES option of R.I.S.K. (in this case, the intermediary products are removed whence the target is built; and when building the package, the binaries are not copied in the tarball, but moved in the tarball). Time: on my old AMD Athlon (686-class), 1250.16 MHz, this takes less than 2 minutes and a half... The resulting package is 8 Mb. But the installation will need a little more, since : 1) The METAFONT plain format is compiled leading to the program that has the right to be called mf(1) since it passes the TRAP test. 2) All the Computer Modern fontes are compiled, using mf(1). 3) The TeX plain format is compiled leading to the program that has the right to be called tex(1) since it passes the TRIP test. Since the work is done for allowing cross-compilation, the binaries are not used on the matrix, but on the target, i.e. at installation time. For the use of the stuff, I will need to put the essential in the document describing the process in the following days. For Plan9, one question: about the group (since for the user it will be ignored). Which group ? I have decided that the stuff will be installed in: - /lib/kertex for machine independant stuff (the macros, but the compiled fontes too). Plan9 users can bind -a to ``redirect'' what will be written where they want. - /$objtype/bin for the binaries. (idem) - /$objtype/lib/kertex/tex/dump/, and /$objtype/lib/kertex/mf/dump/ for the dump of the tables (the compiled version of the macros, that is loaded say when you call virtex(1) mistex(1), this is virtex(1) but it looks to is dump mistex.fmt to load it. The dumps are not machine independant, even not program independant since they depend on options selected at compile time). (idem) To give you a view of what is installed (and possibly where), here is the (generic: for Unices too) map [I need to change root:wheel to variable things) : #== THE MAIN DIR TREE # Everything is relative (for the not man nor bin stuff) to PKGDIR that # one can set as he sees fit. # + d * $PKGDIR/mf root:wheel 755 # This is where we put Computer Modern _sources_ (parameters and drivers # METAFONT's files). # + d * $PKGDIR/mf/cm root:wheel 755 + d * $PKGDIR/mf/mac root:wheel 755 # If you want to ``localize'' messages, look at D.E.K.'s pool stuff. # + d * $PKGDIR/mf/pool root:wheel 755 + d * $PKGDIR/mf/trap root:wheel 755 # Since to create a METAFONT' base file, one dumps, we call it: dump. # But is is machine dependant. # + d * $TARGETBINDIR/lib root:wheel 755 + d * $TARGETBINDIR/lib/mf root:wheel
Re: [9fans] TeX: hurrah!
1. IFAIK? Can't find that anywhere... 2. Is C++ a problem? Not supported by Plan9? K Alexander Sychev santu...@gmail.com 16/04/2010 10:27:36 am On Fri, 16 Apr 2010 18:07:14 +0400, Karljurgen Feuerherm kfeuerh...@wlu.ca wrote: XeTeX/XeLaTeX do this, I believe... Perhaps they can be ported at some point? IFAIK, XeTeX/XeLaTeX based on C++ code. K Karljürgen G. Feuerherm, PhD Department of Archaeology and Classical Studies Wilfrid Laurier University 75 University Avenue West Waterloo, Ontario N2L 3C5 Tel. (519) 884-1970 x3193 Fax (519) 883-0991 (ATTN Arch. Classics) Alexander Sychev santu...@gmail.com 16/04/2010 9:32:04 am Hello! Congratulations! Have you any plans to adapt the TeX for UTF-8 input? On Fri, 16 Apr 2010 15:57:56 +0400, tlaro...@polynum.com mailto:tlaro...@polynum.com tlaro...@polynum.com mailto:tlaro...@polynum.com wrote: So it compiles without ado under Plan9! And it's pure C89 (POSIX is just for the framework, not for the code: I have removed unneeded dependencies). And it's all the latest versions of the programs. So some numbers: - You will need to download a bundle of 4 chunks (I will put all on my website; I think the licence for TeX/METAFONT allows thie): 1) Donald E. Knuth's sources: all the programs; the sources for the Computer Modern Typefaces and the man pages (roff) I have found here and there or reconstructed. = 6 Mb (compressed) 2) kertex_M: these are the tools for the _matrix_ (the vulcan; the node compiling; the cpu) translating from Pascal to C. = 54 kb (compressed) 3) kertex_T: the R.I.S.K. (see below) framework for building TeX/METAFONT, that is the change files, the template Makefiles, the PASCAL-H extensions implemented as a C library. Plus some public domain software: dvips [it is not compiled for now; it's the next in the list] and some auxiliary tools (afm2tfm etc.); MetaPost (not included for now: next after dvips). = 0.4 Mb 4) The R.I.S.K. comp framework (my stuff) that is just a basic POSIX compliant way of configuring, compiling and installing software. Everything is configured via basic parameters file (you should not have to change something here; but it's all sh(1), ed(1), sed(1) and make(1)). = 125kb (compressed; the bigger par in size is D.E. Knuth and Silvio Levy's CWEB [not used here] that I use intensively). So the sources, it is less than 12 Mb uncompressed. And the sources can be (shall be) read-only so you can put them where you want. Space requirements for compilation : you will need about 33 Mb of free space or, at least, 12 Mb of free space if you use the SAVE_SPACE=YES option of R.I.S.K. (in this case, the intermediary products are removed whence the target is built; and when building the package, the binaries are not copied in the tarball, but moved in the tarball). Time: on my old AMD Athlon (686-class), 1250.16 MHz, this takes less than 2 minutes and a half... The resulting package is 8 Mb. But the installation will need a little more, since : 1) The METAFONT plain format is compiled leading to the program that has the right to be called mf(1) since it passes the TRAP test. 2) All the Computer Modern fontes are compiled, using mf(1). 3) The TeX plain format is compiled leading to the program that has the right to be called tex(1) since it passes the TRIP test. Since the work is done for allowing cross-compilation, the binaries are not used on the matrix, but on the target, i.e. at installation time. For the use of the stuff, I will need to put the essential in the document describing the process in the following days. For Plan9, one question: about the group (since for the user it will be ignored). Which group ? I have decided that the stuff will be installed in: - /lib/kertex for machine independant stuff (the macros, but the compiled fontes too). Plan9 users can bind -a to ``redirect'' what will be written where they want. - /$objtype/bin for the binaries. (idem) - /$objtype/lib/kertex/tex/dump/, and /$objtype/lib/kertex/mf/dump/ for the dump of the tables (the compiled version of the macros, that is loaded say when you call virtex(1) mistex(1), this is virtex(1) but it looks to is dump mistex.fmt to load it. The dumps are not machine independant, even not program independant since they depend on options selected at compile time). (idem) To give you a view of what is installed (and possibly where), here is the (generic: for Unices too) map [I need to change root:wheel to variable things) : #== THE MAIN DIR TREE # Everything is relative (for the not man nor bin stuff) to PKGDIR that # one can set as he sees fit. # + d * $PKGDIR/mf root:wheel 755 # This is where we put Computer Modern _sources_ (parameters and drivers # METAFONT's files). # + d * $PKGDIR/mf/cm root:wheel 755 + d * $PKGDIR/mf/mac root:wheel 755 # If you want to ``localize'' messages, look at D.E.K.'s
Re: [9fans] TeX: hurrah!
2. Is C++ a problem? Not supported by Plan9? not supported by the plan 9 compilers. - erik
Re: [9fans] TeX: hurrah!
From: 9fans-boun...@9fans.net [mailto:9fans-boun...@9fans.net] On Behalf Of Karljurgen Feuerherm Sent: Friday, April 16, 2010 12:20 PM To: Fans of the OS Plan 9 from Bell Labs Subject: Re: [9fans] TeX: hurrah! 1. IFAIK? Can't find that anywhere... 2. Is C++ a problem? Not supported by Plan9? Object-Orientation reduces static provability. May be I'm crazy, but I like it when you can prove that a program is only going to do what it was told to do. IFAIK, XeTeX/XeLaTeX based on C++ code. Not all TeX solutions are though.
Re: [9fans] TeX: hurrah!
On Fri, Apr 16, 2010 at 8:47 AM, Patrick Kelly kameo76...@gmail.com wrote: Object-Orientation reduces static provability. True (or true enough)? Not to engender a flame war, but my gut says there must be some Eiffel, Smalltalk, and LISP folk out there who are big on provability, but I can imagine that there's a case out there for saying not all OO implementations are the same. Is this a Gödel question? How do you prove OO reduces static provability? I'm totally OK with a true enough response like the measured complexity introduced makes it more problematic to determine static provability (as I talk out my ass). -Jack
Re: [9fans] TeX: hurrah!
This doesn't make much sense to me. Object-orientation in itself is simply another level of data abstraction. And for the rest, I think provability is more theoretical than practical, other than the most trivial programmes. I'm beginning to get the impression (or perhaps more accurately am increasingly getting the impression) that the plan9 community is reactionary rather than progressive... not a good characteristic if one is trying to make advances in comparison with one's predecessors... K Patrick Kelly kameo76...@gmail.com 16/04/2010 12:47:03 pm Object-Orientation reduces static provability. May be I'm crazy, but I like it when you can prove that a program is only going to do what it was told to do.
Re: [9fans] TeX: hurrah!
I'm beginning to get the impression (or perhaps more accurately am increasingly getting the impression) that the plan9 community is reactionary rather than progressive... not a good characteristic if one is trying to make advances in comparison with one's predecessors... i think the plan 9 community does come off as reactionary. we tend to fight change too often to no effect or declare things as depricated and useless when they are not. i don't think that object-oriented design is one of these. especially as implemented by c++. i think limbo or go's approach to this (regardless of what you think of the langage) is much more sensible. go of course, goes a little further in what you can do with an interface. - erik
Re: [9fans] TeX: hurrah!
One question. Anyone tried get LaTeX on this TeX port? What are the missing pieces to make it run on it? On Fri, Apr 16, 2010 at 7:17 PM, erik quanstrom quans...@quanstro.net wrote: I'm beginning to get the impression (or perhaps more accurately am increasingly getting the impression) that the plan9 community is reactionary rather than progressive... not a good characteristic if one is trying to make advances in comparison with one's predecessors... i think the plan 9 community does come off as reactionary. we tend to fight change too often to no effect or declare things as depricated and useless when they are not. i don't think that object-oriented design is one of these. especially as implemented by c++. i think limbo or go's approach to this (regardless of what you think of the langage) is much more sensible. go of course, goes a little further in what you can do with an interface. - erik
Re: [9fans] TeX: hurrah!
I was just speaking generally. One of my major programming languages is Ada, and I doubt anyone would say that isn't big on provability. I've used objects a couple times, in places where they do in fact help, but those cases are, in general, not read properly. Using an object in the wrong place, which is most places, does lead to worse code. For most people, using the wrong tool for the wrong job is foolish, but for OOP lovers... The question isn't how do you prove it does reduce static provability, but how do you prove it does not. I can cite mathematical proof that the sun revolves around the earth, but we all know that's not true. That being said, there are studies out there about using the wrong paradigm for the wrong job, objects do come up. -Original Message- From: 9fans-boun...@9fans.net [mailto:9fans-boun...@9fans.net] On Behalf Of Jack Johnson Sent: Friday, April 16, 2010 1:05 PM To: Fans of the OS Plan 9 from Bell Labs Subject: Re: [9fans] TeX: hurrah! On Fri, Apr 16, 2010 at 8:47 AM, Patrick Kelly kameo76...@gmail.com wrote: Object-Orientation reduces static provability. True (or true enough)? Not to engender a flame war, but my gut says there must be some Eiffel, Smalltalk, and LISP folk out there who are big on provability, but I can imagine that there's a case out there for saying not all OO implementations are the same. Is this a Gödel question? How do you prove OO reduces static provability? I'm totally OK with a true enough response like the measured complexity introduced makes it more problematic to determine static provability (as I talk out my ass). -Jack
Re: [9fans] TeX: hurrah!
On Fri, Apr 16, 2010 at 10:27 AM, Alexander Sychev santu...@gmail.com wrote: IFAIK, XeTeX/XeLaTeX based on C++ code. XeTeX itself is based on patches to Knuth's WEB source code for TeX. It's the PDF-producing section (xdvipdf or some such) that's written using a C++ library for handling PDF. There will be the same problem with LuaTeX; I think also with pdfTeX. —Joel Salomon
Re: [9fans] TeX: hurrah!
On Fri, Apr 16, 2010 at 07:46:08PM +0200, Francisco J Ballesteros wrote: One question. Anyone tried get LaTeX on this TeX port? What are the missing pieces to make it run on it? I will release/publish the things on Monday I think. In theory, if LaTeX is still a set of macros, it should work with it, following this process (it will be put in the README) : 1) # Dump latex.tex (whether cd to the directory where latex.tex is, or # put latex.tex in TEXINPUTS search paths): # cat EOT | initex \input latex \dump EOT 2) # Put the dump (latex.fmt) in the directory where TeX looks for # dumps (in kerTeX : /$objtype/lib/tex/dump/). # mv latex.fmt /$objtype/lib/tex/dump/ 3) # Call an instance of virtex : latex, so that looking to argv[0], it # loads the corresponding dump: # cd /$objtype/bin cp virtex latex and you will be done... as long as what LaTeX calls is found: other macros (set env TEXINPUTS with the ':' separated directories) and TFM (TEXFONTS). By default, a BIG TeX is generated, that is a binary with extended size tables and so on. If when dumping a package---say LaTeX---it moans capacity exceeded, I need to adjust/increase the values in the change file. --- Thierry Laronde tlaronde +AT+ polynum +dot+ com http://www.kergis.com/ Key fingerprint = 0FF7 E906 FBAF FE95 FD89 250D 52B1 AE95 6006 F40C
Re: [9fans] TeX: hurrah!
From: 9fans-boun...@9fans.net [mailto:9fans-boun...@9fans.net] On Behalf Of Karljurgen Feuerherm Sent: Friday, April 16, 2010 1:11 PM To: 'Fans of the OS Plan 9 from Bell Labs' Subject: Re: [9fans] TeX: hurrah! This doesn't make much sense to me. Object-orientation in itself is simply another level of data abstraction. And for the rest, I think provability is more theoretical than practical, other than the most trivial programmes. You are correct in saying OOP is just another type of abstraction. I would be correct in saying a hammer is just another type of tool, but would you use a hammer to repair circuitry? I would be correct in saying a nuclear missile is just another weapon, but would you use that when a gun would suffice? Was Hiroshima and Nagasaki just simple warfare? There are proper tools for a given job. Using the correct tool is a good thing, using the wrong tool causes problems. There are places where objects are useful, yet most OOP programmers use objects everywhere. When representing non-objects as objects, problems arise. Quite the contrary, provability is most important in the most complex programs. Math is all about proving things, be it simple or complex. If a given formulae is not provable, it's damn near useless, but when a program isn't provable, it's just shrugged off. I'm beginning to get the impression (or perhaps more accurately am increasingly getting the impression) that the plan9 community is reactionary rather than progressive... not a good characteristic if one is trying to make advances in comparison with one's predecessors... Do not let your impression of a few people effect your view of the Plan 9 community. Have you look at what Plan 9 has done? I would hardly go to say we are reactive. Every other system has reacted to what Plan 9 has done, not the other way around. Communities are made up of humans, humans are complex systems. Complex systems cannot simply be labeled reactive or progressive. I'm beginning to get the impression people love to troll Plan 9, more than they love to better computing, more than they love to learn from what Plan 9 has done, more than they love to solve problems, the whole reason behind computer science. K Patrick Kelly kameo76...@gmail.com 16/04/2010 12:47:03 pm Object-Orientation reduces static provability. May be I'm crazy, but I like it when you can prove that a program is only going to do what it was told to do.
Re: [9fans] TeX: hurrah!
Ok--so it's agreed that it's not OO that's the problem, it's the users, then, who don't know which tool to use when. Not at all the same thing. And to be pedantic, since you give this example, the sun does revolve around the earth, so long as you choose the earth as your point of reference... Certain points of reference are to be preferred for certain things, as you said. So OO or not, as appropriate. K Patrick Kelly kameo76...@gmail.com 16/04/2010 1:55:50 pm I was just speaking generally. One of my major programming languages is Ada, and I doubt anyone would say that isn't big on provability. I've used objects a couple times, in places where they do in fact help, but those cases are, in general, not read properly. Using an object in the wrong place, which is most places, does lead to worse code. For most people, using the wrong tool for the wrong job is foolish, but for OOP lovers... The question isn't how do you prove it does reduce static provability, but how do you prove it does not. I can cite mathematical proof that the sun revolves around the earth, but we all know that's not true. That being said, there are studies out there about using the wrong paradigm for the wrong job, objects do come up.
Re: [9fans] TeX: hurrah!
Sorry to be a grouch, but can we change this thread to OO instead of the advertised TeX:hurrah! thread? I'm interested in the TeX news, but not so interested in the OO/language debate that no doubt will go on for a while... Thanks! -joe On Fri, Apr 16, 2010 at 2:14 PM, Karljurgen Feuerherm kfeuerh...@wlu.cawrote: Ok--so it's agreed that it's not OO that's the problem, it's the users, then, who don't know which tool to use when. Not at all the same thing. And to be pedantic, since you give this example, the sun does revolve around the earth, so long as you choose the earth as your point of reference... Certain points of reference are to be preferred for certain things, as you said. So OO or not, as appropriate. K Patrick Kelly kameo76...@gmail.com 16/04/2010 1:55:50 pm I was just speaking generally. One of my major programming languages is Ada, and I doubt anyone would say that isn't big on provability. I've used objects a couple times, in places where they do in fact help, but those cases are, in general, not read properly. Using an object in the wrong place, which is most places, does lead to worse code. For most people, using the wrong tool for the wrong job is foolish, but for OOP lovers... The question isn't how do you prove it does reduce static provability, but how do you prove it does not. I can cite mathematical proof that the sun revolves around the earth, but we all know that's not true. That being said, there are studies out there about using the wrong paradigm for the wrong job, objects do come up.
Re: [9fans] TeX: hurrah!
From: 9fans-boun...@9fans.net [mailto:9fans-boun...@9fans.net] On Behalf Of Karljurgen Feuerherm Sent: Friday, April 16, 2010 2:15 PM To: 'Fans of the OS Plan 9 from Bell Labs' Subject: Re: [9fans] TeX: hurrah! Ok--so it's agreed that it's not OO that's the problem, it's the users, then, who don't know which tool to use when. Not at all the same thing. No, we don't agree. I said there were places objects were useful, I never said object orientation was useful; not at all, the same thing. You can write imperative-style code in a functional language, and you can write functional-style code in an imperative language. That does not mean you code is auto-magically imperative or functional. K
Re: [9fans] TeX: hurrah!
This page and its links maybe be interesting for understanding the relationship between latex and tex: http://www.tug.org/levels.html In my area of computer science all publications are written in latex and for a particular conference/journal a latex class or style file (I must admit to not really knowing what the difference is) is provided and must be adhered to. Everybody I know also used texlive which seems to be the standard tex distribution. It used to be tetex but this is no longer maintained. I would be great to be able to write on plan 9 and I'm very pleased to see the porting effort for tex. James On Fri, Apr 16, 2010 at 9:22 PM, Joseph Stewart joseph.stew...@gmail.com wrote: Sorry to be a grouch, but can we change this thread to OO instead of the advertised TeX:hurrah! thread? I'm interested in the TeX news, but not so interested in the OO/language debate that no doubt will go on for a while... Thanks! -joe On Fri, Apr 16, 2010 at 2:14 PM, Karljurgen Feuerherm kfeuerh...@wlu.ca wrote: Ok--so it's agreed that it's not OO that's the problem, it's the users, then, who don't know which tool to use when. Not at all the same thing. And to be pedantic, since you give this example, the sun does revolve around the earth, so long as you choose the earth as your point of reference... Certain points of reference are to be preferred for certain things, as you said. So OO or not, as appropriate. K Patrick Kelly kameo76...@gmail.com 16/04/2010 1:55:50 pm I was just speaking generally. One of my major programming languages is Ada, and I doubt anyone would say that isn't big on provability. I've used objects a couple times, in places where they do in fact help, but those cases are, in general, not read properly. Using an object in the wrong place, which is most places, does lead to worse code. For most people, using the wrong tool for the wrong job is foolish, but for OOP lovers... The question isn't how do you prove it does reduce static provability, but how do you prove it does not. I can cite mathematical proof that the sun revolves around the earth, but we all know that's not true. That being said, there are studies out there about using the wrong paradigm for the wrong job, objects do come up.
Re: [9fans] TeX: hurrah!
Thanks for this. And yes, indeed, a step in the right direction! Best K James Chapman ja...@cs.ioc.ee 16/04/2010 2:37:20 pm This page and its links maybe be interesting for understanding the relationship between latex and tex: http://www.tug.org/levels.html In my area of computer science all publications are written in latex and for a particular conference/journal a latex class or style file (I must admit to not really knowing what the difference is) is provided and must be adhered to. Everybody I know also used texlive which seems to be the standard tex distribution. It used to be tetex but this is no longer maintained. I would be great to be able to write on plan 9 and I'm very pleased to see the porting effort for tex. James On Fri, Apr 16, 2010 at 9:22 PM, Joseph Stewart joseph.stew...@gmail.com wrote: Sorry to be a grouch, but can we change this thread to OO instead of the advertised TeX:hurrah! thread? I'm interested in the TeX news, but not so interested in the OO/language debate that no doubt will go on for a while... Thanks! -joe On Fri, Apr 16, 2010 at 2:14 PM, Karljurgen Feuerherm kfeuerh...@wlu.ca wrote: Ok--so it's agreed that it's not OO that's the problem, it's the users, then, who don't know which tool to use when. Not at all the same thing. And to be pedantic, since you give this example, the sun does revolve around the earth, so long as you choose the earth as your point of reference... Certain points of reference are to be preferred for certain things, as you said. So OO or not, as appropriate. K Patrick Kelly kameo76...@gmail.com 16/04/2010 1:55:50 pm I was just speaking generally. One of my major programming languages is Ada, and I doubt anyone would say that isn't big on provability. I've used objects a couple times, in places where they do in fact help, but those cases are, in general, not read properly. Using an object in the wrong place, which is most places, does lead to worse code. For most people, using the wrong tool for the wrong job is foolish, but for OOP lovers... The question isn't how do you prove it does reduce static provability, but how do you prove it does not. I can cite mathematical proof that the sun revolves around the earth, but we all know that's not true. That being said, there are studies out there about using the wrong paradigm for the wrong job, objects do come up.
Re: [9fans] TeX: hurrah!
It occurred to me that a profitable thing to do here would be to mention some things that would be nice to see in a new improved TeX... I believe bidirectional was mentioned already. The other thing that is essential for folk like me is complete Unicode compatibility [Yes, I know. UTC has committed many sins :] BUT: just so it's clear to everyone that I'm not just trying to get goodies on the backs of other people's hard work--I'd be more than willing to help in due course. First I have to get a Plan9 installation, though, and I need to acknowledge that I'm relatively unfamiliar with the guts of TeX at the present. So one step at at time :) K
Re: [9fans] TeX: hurrah!
On Fri, Apr 16, 2010 at 6:10 PM, Karljurgen Feuerherm kfeuerh...@wlu.ca wrote: It occurred to me that a profitable thing to do here would be to mention some things that would be nice to see in a new improved TeX... I believe bidirectional was mentioned already. The other thing that is essential for folk like me is complete Unicode compatibility [Yes, I know. UTC has committed many sins :] LaTeX has the inputenc package. BUT: just so it's clear to everyone that I'm not just trying to get goodies on the backs of other people's hard work--I'd be more than willing to help in due course. First I have to get a Plan9 installation, though, and I need to acknowledge that I'm relatively unfamiliar with the guts of TeX at the present. So one step at at time :) K
