[sage-support] Re: losing confidence in Sage notebooks
One thing you could do is submit bug reports, especially with reproducible answers. If I could, I would. Unfortunately, these are rare events, they just have a big impact when they happen. They may be related to crashes, unclean shutdowns, browser crashes, etc. Many variables and rare events means it's hard to figure out how to reproduce them. Tom -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: losing confidence in Sage notebooks
You can put the .sage/sage_notebook.sagenb/home/ directory under revision control. You should probably only hg add the worksheet.html and worksheet_conf.pickle files, and ignore everything else. That doesn't really work with the current directory structure. For example, imagine I add two different worksheets on two systems and then try to synchronize--the directory names will conflict even though the worksheets have different names. Conversely, the same worksheet may be in two different directories, so it would never get merged. Also, I would need to manually add and commit things from the command line. So, using version control right now is at least as much work as doing it by hand, and it makes merging and recovering from problems even more complex than it already is. The problem is really with the .sage directory structure; without changing that, I think it will be hard for Sage to get reliable version control or error recovery. Fortunately, the changes may not have to be big. Right now, .sage is essentially an opaque directory that I can't reliably do anything with. You as a developer may know what's in there and how it works, but I don't. Look at the path: .sage/sage_notebook/worksheets/admin/10 So, here are some sample questions that come up: -- What worksheet is that? (I need to look inside--makes it hard to organize manually.) -- How do I copy a new worksheet into that directory from somewhere else? -- Is the title stored anywhere else? Which of the files contains the authoritative title? -- Is anything cached? On disk? In memory? When does the cache get updated? -- Can I just edit worksheet.txt? -- What happens if I make changes while the server is running, will it blow those away? -- If I copy in a different worksheet.txt in there, do things just work? Do I need to delete something else? I can read a bit between the lines if I go into the Python API documentation for sage.server.notebook.worksheet.Worksheet. But the existence of a programmatic API like that may even suggest that I should make no assumptions about the representation of worksheets at all. My suggestion would be the following: -- For single user mode, put all the worksheets in ~/SageMath (that tells me that I'm supposed to look at them) -- Put all notebook files at the toplevel with descriptive names, like ~/SageMath/fft.sws -- Put all cache/computed files in ~/SageMath/.cache/... -- Guarantee the user that he can change anything in ~/SageMath (outside .cache) by hand and things will keep working; the server will detect changes and rebuild cache files as necessary. -- Adding a new worksheet is the same as copying a file into ~/ SageMath/foo.sws -- Put in hooks to trigger hg add, hg commit, etc. when necessary (create a .hgignore containing .cache automatically) -- Trigger commits fairly frequently (they're cheap) With a directory structure like that, things get a lot simpler for me because I actually understand what I can do: -- Make a backup: cp ~/SageMath/*.sws ~/my-sage-backup -- Restore backup: cp ~/my-sage-backup/* ~/SageMath -- Restore a few files: cp ~/my-sage-backup/improc*.sws ~/SageMath -- Fix something in the fft worksheet: vi ~/SageMath/fft.sws -- Duplicate a worksheet: cp ~/SageMath/fft.sws ~/SageMath/new-fft.sws -- Synchronize repositories: cd ~/SageMath; hg pull ssh://.../SageMath; hg update -- Merging of worksheets in version control actually works as expected. I understand that somehow I can probably do many of these things already. But the structure of the current directory tree doesn't make it easy or even easy to figure out, and the documentation has not been that much help to me either. With a clearer directory tree layout and some small guarantees about the cache, you wouldn't need much documentation for people to be able to work with sage worksheets at the directory level reliably. And, in particular, it would make it much easier for people like me to recover from data loss due to other bugs in Sage. Tom PS: For multiuser using the current setup, you could use ~/SageMath for admin, and ~/SageUsers/user-name/... for all other user accounts. (Of course, in a university environment, we'd really like to use LDAP for authentication and ~user/SageMath for each valid user.) -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: large upload fails
My desktop machine, an 8G 4-core machine. I've tried multiple browsers. This is not a problem with the browser; browsers are capable of very large and very slow uploads. Probably some timeout parameter in the Sage server is set too small. There is no output related to the upload on the Sage console. Tom On Dec 4, 3:56 am, William Stein wst...@gmail.com wrote: 2009/12/3 Robert Bradshaw rober...@math.washington.edu: On Dec 3, 2009, at 4:51 PM, tmbdev wrote: I backed up my worksheets in a download_worksheets.zip file. The file is 63M large. Now I'm trying to upload it and the upload is failing. Is there any way of getting the zip file uploaded or installed (maybe I can just unzip it somewhere)? What server are you uploading them into? sagenb.org? Your desktop? -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: losing confidence in Sage notebooks
Note, incidentally, that restoring from backup (whether in parts or not) is a major problem because the paths to the notebooks change. So, links that I put into the lecture notes point to the wrong notebooks, or are just dead. That's another reason why the current naming scheme really needs to change somehow. Tom -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
just a side remark - IMHO notebooks are not designed for any kind of large-scope project. For the latter, you are much better off with good old scripts. Notebooks certainly have their own pluses, such as more interactivity and ease of collaboration---but not sharing, as was demonstrated here recently. Otherwise I don't really see a point of them. Best, Dmitrii 2009/12/4 tmb tmb...@gmail.com: Note, incidentally, that restoring from backup (whether in parts or not) is a major problem because the paths to the notebooks change. So, links that I put into the lecture notes point to the wrong notebooks, or are just dead. That's another reason why the current naming scheme really needs to change somehow. -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] 3D plots fail using VirtualBox Sage 4.21 on old XP computer
I am running Sage 4.2.1 on a old 1GB ram XP computer with the VirtualBox binary distribution. In the notebook if I try a 3D graph Firefox shuts down and kills the Sage session. This is the same problem I had with the VirtualBox Sage version 4.2. I have also created a separate Ubuntu 9.1 vm in VirtualBox, where I loaded all the stuff needed to run and compile Sage 4.21. This included the FireFox upgrades needed to run the 3D plots from a notebook. To this vm I added a second network connection which is a VirtualBox host only, just like in the binary VirtualBox Sage 4.21. I compiled Sage 4.21. When I run the notebook sage using the Firefox inside the Ubuntu 9.1 vm everything works great and I get 3D graphs. I can also start a Sage notebook with a IP address and access it from my Windows Firefox. This works fine until I run a 3D graph and then the windows Firefox closes down, just like in the binary distribution for VirtualBox Sage 4.21. However, if I start Firefox in Ubuntu I can get into the same sheet and run the 3D graph with no problem and Sage has not shut down. I suspect the problem with the 3D graph is from the VirtualBox host connection. I don't know if it is just because of my XP system. In general I like the VirtualBox system better than VMware. However, for users with old XP systems it is a problem. Below I included my system specs. Mike Intel Celeron M 1.5 GHz, 1 GB Ram Microsoft Windows XP Professional 5.1.2600, Service Pack 3 VirtualBox 3.0.10 results from cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model: 13 model name : Intel(R) Celeron(R) M processor 1.50GHz stepping: 8 cpu MHz : 1476.382 cache size: 1024 KB fdiv_bug: no hlt_bug : no f00f_bug: no coma_bug : no fpu: yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss nx up bogomips : 3007.50 -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] 3D plots fail using VirtualBox Sage 4.21 on old XP computer
Michael, I guess you have discovered (or about to discover) yourself a well–known fact that Windows (XP, or whatever) is not really suitable platform for doing any remotely serious computing :) Dmitrii 2009/12/4 Michael Madison madison.mich...@gmail.com: I am running Sage 4.2.1 on a old 1GB ram XP computer with the VirtualBox binary distribution. In the notebook if I try a 3D graph Firefox shuts down and kills the Sage session. This is the same problem I had with the VirtualBox Sage version 4.2. I have also created a separate Ubuntu 9.1 vm in VirtualBox, where I loaded all the stuff needed to run and compile Sage 4.21. This included the FireFox upgrades needed to run the 3D plots from a notebook. To this vm I added a second network connection which is a VirtualBox host only, just like in the binary VirtualBox Sage 4.21. I compiled Sage 4.21. When I run the notebook sage using the Firefox inside the Ubuntu 9.1 vm everything works great and I get 3D graphs. I can also start a Sage notebook with a IP address and access it from my Windows Firefox. This works fine until I run a 3D graph and then the windows Firefox closes down, just like in the binary distribution for VirtualBox Sage 4.21. However, if I start Firefox in Ubuntu I can get into the same sheet and run the 3D graph with no problem and Sage has not shut down. I suspect the problem with the 3D graph is from the VirtualBox host connection. I don't know if it is just because of my XP system. In general I like the VirtualBox system better than VMware. However, for users with old XP systems it is a problem. Below I included my system specs. Mike Intel Celeron M 1.5 GHz, 1 GB Ram Microsoft Windows XP Professional 5.1.2600, Service Pack 3 VirtualBox 3.0.10 results from cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Celeron(R) M processor 1.50GHz stepping : 8 cpu MHz : 1476.382 cache size : 1024 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss nx up bogomips : 3007.50 -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org -- Dmitrii Pasechnik - DISCLAIMER: Any text following this sentence does not constitute a part of this message, and was added automatically during transmission. -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: losing confidence in Sage notebooks
On Dec 4, 1:53 pm, Dima Pasechnik dimp...@gmail.com wrote: just a side remark - IMHO notebooks are not designed for any kind of large-scope project. I'm not using them for a large-scope project, I'm using them for teaching pattern recognition and image processing. That means that I create lots of simple notebooks, a few for each lecture, published so that students can look at them, copy them, and experiment. That seems to me like a primary use case for Sage. By and large, notebooks are good for that, they're just a bit buggy and hard to back up/restore. I'm not even attempting to use it for my research yet. For that, there really are some features missing. Tom -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
Tom, I'm contemplating using Sage for teaching an OR course next term, but I won't touch notebooks. I'd just use plain-text scripts. Yes, this would require having Sage installed on a server, where they would be able to run usual interactive sessions. Nothing is wrong with such setup, IMHO... Dmitrii 2009/12/4 tmb tmb...@gmail.com: On Dec 4, 1:53 pm, Dima Pasechnik dimp...@gmail.com wrote: just a side remark - IMHO notebooks are not designed for any kind of large-scope project. I'm not using them for a large-scope project, I'm using them for teaching pattern recognition and image processing. That means that I create lots of simple notebooks, a few for each lecture, published so that students can look at them, copy them, and experiment. That seems to me like a primary use case for Sage. By and large, notebooks are good for that, they're just a bit buggy and hard to back up/restore. I'm not even attempting to use it for my research yet. For that, there really are some features missing. Tom -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org -- Dmitrii Pasechnik - DISCLAIMER: Any text following this sentence does not constitute a part of this message, and was added automatically during transmission. -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: 3D plots fail using VirtualBox Sage 4.21 on old XP computer
I think we should say then what base system requirements are to run Sage. However, running Sage in a VirtualBox Ubuntu 9.1 with the Ubuntu Firefox works fine. I have new system coming. I just wanted to point out if Sage has users trying to run a system like mine it has problems. Thanks Mike On Dec 4, 7:29 am, Dima Pasechnik dimp...@gmail.com wrote: Michael, I guess you have discovered (or about to discover) yourself a well–known fact that Windows (XP, or whatever) is not really suitable platform for doing any remotely serious computing :) Dmitrii 2009/12/4 Michael Madison madison.mich...@gmail.com: I am running Sage 4.2.1 on a old 1GB ram XP computer with the VirtualBox binary distribution. In the notebook if I try a 3D graph Firefox shuts down and kills the Sage session. This is the same problem I had with the VirtualBox Sage version 4.2. I have also created a separate Ubuntu 9.1 vm in VirtualBox, where I loaded all the stuff needed to run and compile Sage 4.21. This included the FireFox upgrades needed to run the 3D plots from a notebook. To this vm I added a second network connection which is a VirtualBox host only, just like in the binary VirtualBox Sage 4.21. I compiled Sage 4.21. When I run the notebook sage using the Firefox inside the Ubuntu 9.1 vm everything works great and I get 3D graphs. I can also start a Sage notebook with a IP address and access it from my Windows Firefox. This works fine until I run a 3D graph and then the windows Firefox closes down, just like in the binary distribution for VirtualBox Sage 4.21. However, if I start Firefox in Ubuntu I can get into the same sheet and run the 3D graph with no problem and Sage has not shut down. I suspect the problem with the 3D graph is from the VirtualBox host connection. I don't know if it is just because of my XP system. In general I like the VirtualBox system better than VMware. However, for users with old XP systems it is a problem. Below I included my system specs. Mike Intel Celeron M 1.5 GHz, 1 GB Ram Microsoft Windows XP Professional 5.1.2600, Service Pack 3 VirtualBox 3.0.10 results from cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Celeron(R) M processor 1.50GHz stepping : 8 cpu MHz : 1476.382 cache size : 1024 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss nx up bogomips : 3007.50 -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group athttp://groups.google.com/group/sage-support URL:http://www.sagemath.org -- Dmitrii Pasechnik - DISCLAIMER: Any text following this sentence does not constitute a part of this message, and was added automatically during transmission.- Hide quoted text - - Show quoted text - -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: 3D plots fail using VirtualBox Sage 4.21 on old XP computer
Though Windows is not near as well supported as linux and OS X, this should work in principle. Try browsing some of the applets at http://jmol.sourceforge.net/ , does that crash your browser? - Robert On Dec 4, 2009, at 8:09 AM, Michael Madison wrote: I think we should say then what base system requirements are to run Sage. However, running Sage in a VirtualBox Ubuntu 9.1 with the Ubuntu Firefox works fine. I have new system coming. I just wanted to point out if Sage has users trying to run a system like mine it has problems. Thanks Mike On Dec 4, 7:29 am, Dima Pasechnik dimp...@gmail.com wrote: Michael, I guess you have discovered (or about to discover) yourself a well–known fact that Windows (XP, or whatever) is not really suitable platform for doing any remotely serious computing :) Dmitrii 2009/12/4 Michael Madison madison.mich...@gmail.com: I am running Sage 4.2.1 on a old 1GB ram XP computer with the VirtualBox binary distribution. In the notebook if I try a 3D graph Firefox shuts down and kills the Sage session. This is the same problem I had with the VirtualBox Sage version 4.2. I have also created a separate Ubuntu 9.1 vm in VirtualBox, where I loaded all the stuff needed to run and compile Sage 4.21. This included the FireFox upgrades needed to run the 3D plots from a notebook. To this vm I added a second network connection which is a VirtualBox host only, just like in the binary VirtualBox Sage 4.21. I compiled Sage 4.21. When I run the notebook sage using the Firefox inside the Ubuntu 9.1 vm everything works great and I get 3D graphs. I can also start a Sage notebook with a IP address and access it from my Windows Firefox. This works fine until I run a 3D graph and then the windows Firefox closes down, just like in the binary distribution for VirtualBox Sage 4.21. However, if I start Firefox in Ubuntu I can get into the same sheet and run the 3D graph with no problem and Sage has not shut down. I suspect the problem with the 3D graph is from the VirtualBox host connection. I don't know if it is just because of my XP system. In general I like the VirtualBox system better than VMware. However, for users with old XP systems it is a problem. Below I included my system specs. Mike Intel Celeron M 1.5 GHz, 1 GB Ram Microsoft Windows XP Professional 5.1.2600, Service Pack 3 VirtualBox 3.0.10 results from cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model: 13 model name : Intel(R) Celeron(R) M processor 1.50GHz stepping: 8 cpu MHz : 1476.382 cache size: 1024 KB fdiv_bug: no hlt_bug : no f00f_bug: no coma_bug : no fpu: yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss nx up bogomips : 3007.50 -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group athttp://groups.google.com/group/sage-support URL:http://www.sagemath.org -- Dmitrii Pasechnik - DISCLAIMER: Any text following this sentence does not constitute a part of this message, and was added automatically during transmission.- Hide quoted text - - Show quoted text - -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] resolution of issue with Tkinter, sage-4.2.1, and SLED 10.2
All: I reported an issue a few days ago in which sage's python 2.6 was not properly configured for tcl/tk. My native stand-alone version of python 2.6 (outside of sage) was just fine,though. I've resolved the issue, thanks to a suggestion from this discussion group. I went to my system-wide installation of tcl/tk and did the following: 1. Copied the tcl and tk header files to my sage-local/sage-4.2.1/ local/include directory. The above is in the directory structure where I am doing the sage build. 2 Copied the tcl and tk libraries to my sage-local/sage-4.2.1/local/ lib directory. 3. re-installed sage's python 2.6 build via the command ./sage -f python-2.6.2.p4 I can now import Tkinter inside the sage shell, and executing the command Tkinter._test() pops open a gui window with two buttons, as it should. This behavior is now consistent with my stand-alone python 2.6 install. Thanks for the help. T. Davis -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] 3D plots fail using VirtualBox Sage 4.21 on old XP computer
Hi, 1) Try reducing the amount of memory allocated to the Sage Virtual machine from 512MB to 384MB. 2) Windows XP (even with only 1GB) is a very important platform for Sage, IMHO. -- William On Fri, Dec 4, 2009 at 10:21 AM, Michael Madison madison.mich...@gmail.com wrote: I am running Sage 4.2.1 on a old 1GB ram XP computer with the VirtualBox binary distribution. In the notebook if I try a 3D graph Firefox shuts down and kills the Sage session. This is the same problem I had with the VirtualBox Sage version 4.2. I have also created a separate Ubuntu 9.1 vm in VirtualBox, where I loaded all the stuff needed to run and compile Sage 4.21. This included the FireFox upgrades needed to run the 3D plots from a notebook. To this vm I added a second network connection which is a VirtualBox host only, just like in the binary VirtualBox Sage 4.21. I compiled Sage 4.21. When I run the notebook sage using the Firefox inside the Ubuntu 9.1 vm everything works great and I get 3D graphs. I can also start a Sage notebook with a IP address and access it from my Windows Firefox. This works fine until I run a 3D graph and then the windows Firefox closes down, just like in the binary distribution for VirtualBox Sage 4.21. However, if I start Firefox in Ubuntu I can get into the same sheet and run the 3D graph with no problem and Sage has not shut down. I suspect the problem with the 3D graph is from the VirtualBox host connection. I don't know if it is just because of my XP system. In general I like the VirtualBox system better than VMware. However, for users with old XP systems it is a problem. Below I included my system specs. Mike Intel Celeron M 1.5 GHz, 1 GB Ram Microsoft Windows XP Professional 5.1.2600, Service Pack 3 VirtualBox 3.0.10 results from cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Celeron(R) M processor 1.50GHz stepping : 8 cpu MHz : 1476.382 cache size : 1024 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss nx up bogomips : 3007.50 -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org -- William Stein Associate Professor of Mathematics University of Washington http://wstein.org -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: losing confidence in Sage notebooks
I'm not even attempting to use it for my research yet. For that, there really are some features missing. Incidentally, please feel free to be explicit about what features are missing for you to use it for research - and, if possible, any open source projects which *do* have such features and would be possible to add to Sage, especially Python, C, or C++ libraries. - kcrisman -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
On Dec 4, 2009, at 4:53 AM, Dima Pasechnik wrote: just a side remark - IMHO notebooks are not designed for any kind of large-scope project. There's no reason they couldn't be. For the latter, you are much better off with good old scripts. Notebooks certainly have their own pluses, such as more interactivity and ease of collaboration---but not sharing, as was demonstrated here recently. Otherwise I don't really see a point of them. Ease of use, especially for people on Windows. I think they have advantages for sharing as well, as I can share a notebook with someone by posting it on a public (or personal) server, and they can go and get an account and use it without having to even install sage. This could be particularly nice for the classroom setting. 2009/12/4 tmb tmb...@gmail.com: Note, incidentally, that restoring from backup (whether in parts or not) is a major problem because the paths to the notebooks change. So, links that I put into the lecture notes point to the wrong notebooks, or are just dead. That's another reason why the current naming scheme really needs to change somehow. I agree. It would make for longer URLs and paths, but maybe we should have globally unique, and consistent, ids for worksheets (for example, just choosing a sufficiently large random number and using the ostrich algorithm to resolve collisions). Then one could at least link between worksheets reliably, and correlate worksheets across multiple servers. - Robert -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
On Dec 4, 2009, at 3:39 AM, tmb wrote: You can put the .sage/sage_notebook.sagenb/home/ directory under revision control. You should probably only hg add the worksheet.html and worksheet_conf.pickle files, and ignore everything else. That doesn't really work with the current directory structure. For example, imagine I add two different worksheets on two systems and then try to synchronize--the directory names will conflict even though the worksheets have different names. Conversely, the same worksheet may be in two different directories, so it would never get merged. Also, I would need to manually add and commit things from the command line. So, using version control right now is at least as much work as doing it by hand, and it makes merging and recovering from problems even more complex than it already is. The problem is really with the .sage directory structure; without changing that, I think it will be hard for Sage to get reliable version control or error recovery. Fortunately, the changes may not have to be big. Right now, .sage is essentially an opaque directory that I can't reliably do anything with. You as a developer may know what's in there and how it works, but I don't. Most of what's there is caching, temporary files, user prefs ,etc. (More than just the notebook, btw.) Look at the path: .sage/sage_notebook/worksheets/admin/10 So, here are some sample questions that come up: -- What worksheet is that? (I need to look inside--makes it hard to organize manually.) -- How do I copy a new worksheet into that directory from somewhere else? -- Is the title stored anywhere else? Which of the files contains the authoritative title? -- Is anything cached? On disk? In memory? When does the cache get updated? -- Can I just edit worksheet.txt? -- What happens if I make changes while the server is running, will it blow those away? -- If I copy in a different worksheet.txt in there, do things just work? Do I need to delete something else? I don't know the answer to all those questions, but a little experimentation might shed some light on it (as would reading the code as you mention.) It is accurate to say the current structure is a implementation detail, not a promised API. I can read a bit between the lines if I go into the Python API documentation for sage.server.notebook.worksheet.Worksheet. But the existence of a programmatic API like that may even suggest that I should make no assumptions about the representation of worksheets at all. My suggestion would be the following: -- For single user mode, put all the worksheets in ~/SageMath (that tells me that I'm supposed to look at them) It's standard to make a .foo directory to hold application defaults and data. Making a visible top-level directory is more invasive. Not treating admin specially also makes the code cleaner. -- Put all notebook files at the toplevel with descriptive names, like ~/SageMath/fft.sws Where would the descriptive names come from? How would one handle naming conflicts? Worksheet renaming? -- Put all cache/computed files in ~/SageMath/.cache/... -- Guarantee the user that he can change anything in ~/SageMath (outside .cache) by hand and things will keep working; the server will detect changes and rebuild cache files as necessary. -- Adding a new worksheet is the same as copying a file into ~/ SageMath/foo.sws -- Put in hooks to trigger hg add, hg commit, etc. when necessary (create a .hgignore containing .cache automatically) -- Trigger commits fairly frequently (they're cheap) With a directory structure like that, things get a lot simpler for me because I actually understand what I can do: -- Make a backup: cp ~/SageMath/*.sws ~/my-sage-backup -- Restore backup: cp ~/my-sage-backup/* ~/SageMath This works already (with ~/.sage/sage_notebook) -- Restore a few files: cp ~/my-sage-backup/improc*.sws ~/SageMath -- Fix something in the fft worksheet: vi ~/SageMath/fft.sws Note a .sws file may is actually a archive of several relevant files (e.g. images that have been uploaded into DATA). -- Duplicate a worksheet: cp ~/SageMath/fft.sws ~/SageMath/new-fft.sws -- Synchronize repositories: cd ~/SageMath; hg pull ssh://.../ SageMath; hg update -- Merging of worksheets in version control actually works as expected. One still has the issue of naming conflicts between multiple systems. I think it makes more sense to put a repository inside each worksheet. Otherwise two servers have to be totally in sync (exactly the same set of worksheets) or not syncable at all. It would still be easy to write a script to sync every worksheet (say, for a given user) one at a time. But the idea of putting this data under revision control is a good one--no one's had time to implement it yet. Do you want to help out? I understand that somehow I can probably do many of these things already. But the structure of the current
[sage-support] Re: losing confidence in Sage notebooks
-- For single user mode, put all the worksheets in ~/SageMath (that tells me that I'm supposed to look at them) It's standard to make a .foo directory to hold application defaults and data. Making a visible top-level directory is more invasive. Dotfiles contain application defaults, they do not usually contain application data. Application data is usually given on the command line or stored in a toplevel directory. Many IDEs, for example, create toplevel directories in the home directory that contain their projects (this is often configurable in a dotfile). -- Put all notebook files at the toplevel with descriptive names, like ~/SageMath/fft.sws Where would the descriptive names come from? How would one handle naming conflicts? Worksheet renaming? The usual thing to do is that documents have a short name that is used in the file system and URL, and a title that is contained in the document and stored in the index. Google Sites uses that for example. The short name usually remains fixed (but can be changed if necessary) and is used for things like merging versions etc. The title can be changed pretty freely. The short name is initially suggested based on the document title. With a directory structure like that, things get a lot simpler for me because I actually understand what I can do: -- Make a backup: cp ~/SageMath/*.sws ~/my-sage-backup -- Restore backup: cp ~/my-sage-backup/* ~/SageMath This works already (with ~/.sage/sage_notebook) Oh? Why don't you try it. cp ~/SageMath/*.sws ~/my-sage-backup doesn't back up my Sage worksheets. A recursive directory copy of the entire tree does, but that's not particularly interesting. Even there, I don't know whether I can safely restore it on another instance. -- Restore a few files: cp ~/my-sage-backup/improc*.sws ~/SageMath -- Fix something in the fft worksheet: vi ~/SageMath/fft.sws Note a .sws file may is actually a archive of several relevant files (e.g. images that have been uploaded into DATA). Sorry, I meant fft.txt (replace .sws with .txt in all the examples). One still has the issue of naming conflicts between multiple systems. Not if worksheets use explicit file names (see above). I think it makes more sense to put a repository inside each worksheet. That's not convenient when I have 100 or 200 worksheets. It would still be easy to write a script to sync every worksheet (say, for a given user) one at a time. But the idea of putting this data under revision control is a good one--no one's had time to implement it yet. Do you want to help out? Version control and more sensible naming isn't something I need very much in and of itself; if Sage worksheets worked correctly, I'd mostly be happy with it the way it is. And it is not particularly important to me at the individual worksheet level, since most of the problems I'm encountering seem to be happening at the directory tree level (I forgot to mention duplication of worksheets). I'll probably just work some separate Python scripts for copying and merging notebook directories and keep my fingers crossed that the Sage worksheet issues get fixed. Tom -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] 3D plots fail using VirtualBox Sage 4.21 on old XP computer
As reported here in other threads it seems that jmol-based 3-graphics fails in FireFox under Windows XP. I am not sure if it is all configurations anv versions or only some but I do have one laptop running Windows XP and the most recent version of FireFox and this still fails with the newest version of Sage on a remote server. If this is the same problem then I am quite sure it is not a virtual machine problem. It was my understanding that someone (the original jmol developer?) was aware of and working on this problem. No? On Fri, Dec 4, 2009 at 11:59 AM, William Stein wst...@gmail.com wrote: Hi, 1) Try reducing the amount of memory allocated to the Sage Virtual machine from 512MB to 384MB. 2) Windows XP (even with only 1GB) is a very important platform for Sage, IMHO. -- William On Fri, Dec 4, 2009 at 10:21 AM, Michael Madison madison.mich...@gmail.com wrote: I am running Sage 4.2.1 on a old 1GB ram XP computer with the VirtualBox binary distribution. In the notebook if I try a 3D graph Firefox shuts down and kills the Sage session. This is the same problem I had with the VirtualBox Sage version 4.2. I have also created a separate Ubuntu 9.1 vm in VirtualBox, where I loaded all the stuff needed to run and compile Sage 4.21. This included the FireFox upgrades needed to run the 3D plots from a notebook. To this vm I added a second network connection which is a VirtualBox host only, just like in the binary VirtualBox Sage 4.21. I compiled Sage 4.21. When I run the notebook sage using the Firefox inside the Ubuntu 9.1 vm everything works great and I get 3D graphs. I can also start a Sage notebook with a IP address and access it from my Windows Firefox. This works fine until I run a 3D graph and then the windows Firefox closes down, just like in the binary distribution for VirtualBox Sage 4.21. However, if I start Firefox in Ubuntu I can get into the same sheet and run the 3D graph with no problem and Sage has not shut down. I suspect the problem with the 3D graph is from the VirtualBox host connection. I don't know if it is just because of my XP system. -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: losing confidence in Sage notebooks
Incidentally, please feel free to be explicit about what features are missing for you to use it for research - and, if possible, any open source projects which *do* have such features and would be possible to add to Sage, especially Python, C, or C++ libraries. My research involves numerical algorithms and large datasets; so what I'd mostly need is facilities for referencing other notebooks as libraries, facilities for managing and sharing large datasets between notebooks, and somewhat better plotting/graphics facilities. What would also be useful is a facility that triggers recomputation of notebooks automatically (i.e., without loading them manually and selecting recomputation) and notifies me of progress in recomputations. Tom -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] 3D plots fail using VirtualBox Sage 4.21 on old XP computer
William, I am traveling. However, My Ubuntu vm is 380MB. Mike Sent from my iPod On Dec 4, 2009, at 8:59 AM, William Stein wst...@gmail.com wrote: Hi, 1) Try reducing the amount of memory allocated to the Sage Virtual machine from 512MB to 384MB. 2) Windows XP (even with only 1GB) is a very important platform for Sage, IMHO. -- William On Fri, Dec 4, 2009 at 10:21 AM, Michael Madison madison.mich...@gmail.com wrote: I am running Sage 4.2.1 on a old 1GB ram XP computer with the VirtualBox binary distribution. In the notebook if I try a 3D graph Firefox shuts down and kills the Sage session. This is the same problem I had with the VirtualBox Sage version 4.2. I have also created a separate Ubuntu 9.1 vm in VirtualBox, where I loaded all the stuff needed to run and compile Sage 4.21. This included the FireFox upgrades needed to run the 3D plots from a notebook. To this vm I added a second network connection which is a VirtualBox host only, just like in the binary VirtualBox Sage 4.21. I compiled Sage 4.21. When I run the notebook sage using the Firefox inside the Ubuntu 9.1 vm everything works great and I get 3D graphs. I can also start a Sage notebook with a IP address and access it from my Windows Firefox. This works fine until I run a 3D graph and then the windows Firefox closes down, just like in the binary distribution for VirtualBox Sage 4.21. However, if I start Firefox in Ubuntu I can get into the same sheet and run the 3D graph with no problem and Sage has not shut down. I suspect the problem with the 3D graph is from the VirtualBox host connection. I don't know if it is just because of my XP system. In general I like the VirtualBox system better than VMware. However, for users with old XP systems it is a problem. Below I included my system specs. Mike Intel Celeron M 1.5 GHz, 1 GB Ram Microsoft Windows XP Professional 5.1.2600, Service Pack 3 VirtualBox 3.0.10 results from cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model: 13 model name : Intel(R) Celeron(R) M processor 1.50GHz stepping: 8 cpu MHz : 1476.382 cache size: 1024 KB fdiv_bug: no hlt_bug : no f00f_bug: no coma_bug : no fpu: yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss nx up bogomips : 3007.50 -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org -- William Stein Associate Professor of Mathematics University of Washington http://wstein.org -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: 3D plots fail using VirtualBox Sage 4.21 on old XP computer
Robert, Yes, In both my FireFox and Windows Explorer the jmol web page causes both browsers to fail. I also tried to view the 3D plot from Sage in Windows Explorer and got the same failure. So it probably is not sage, but jmol in XP. Again, one way out is running FireFox from within Ubuntu. For general windows users setting up Ubuntu may be a little complicated. I recorded most of the steps so I could post a step by step guide. Mike On Dec 4, 8:24 am, Robert Bradshaw rober...@math.washington.edu wrote: Though Windows is not near as well supported as linux and OS X, this should work in principle. Try browsing some of the applets athttp://jmol.sourceforge.net/ , does that crash your browser? - Robert On Dec 4, 2009, at 8:09 AM, Michael Madison wrote: I think we should say then what base system requirements are to run Sage. However, running Sage in a VirtualBox Ubuntu 9.1 with the Ubuntu Firefox works fine. I have new system coming. I just wanted to point out if Sage has users trying to run a system like mine it has problems. Thanks Mike On Dec 4, 7:29 am, Dima Pasechnik dimp...@gmail.com wrote: Michael, I guess you have discovered (or about to discover) yourself a well–known fact that Windows (XP, or whatever) is not really suitable platform for doing any remotely serious computing :) Dmitrii 2009/12/4 Michael Madison madison.mich...@gmail.com: I am running Sage 4.2.1 on a old 1GB ram XP computer with the VirtualBox binary distribution. In the notebook if I try a 3D graph Firefox shuts down and kills the Sage session. This is the same problem I had with the VirtualBox Sage version 4.2. I have also created a separate Ubuntu 9.1 vm in VirtualBox, where I loaded all the stuff needed to run and compile Sage 4.21. This included the FireFox upgrades needed to run the 3D plots from a notebook. To this vm I added a second network connection which is a VirtualBox host only, just like in the binary VirtualBox Sage 4.21. I compiled Sage 4.21. When I run the notebook sage using the Firefox inside the Ubuntu 9.1 vm everything works great and I get 3D graphs. I can also start a Sage notebook with a IP address and access it from my Windows Firefox. This works fine until I run a 3D graph and then the windows Firefox closes down, just like in the binary distribution for VirtualBox Sage 4.21. However, if I start Firefox in Ubuntu I can get into the same sheet and run the 3D graph with no problem and Sage has not shut down. I suspect the problem with the 3D graph is from the VirtualBox host connection. I don't know if it is just because of my XP system. In general I like the VirtualBox system better than VMware. However, for users with old XP systems it is a problem. Below I included my system specs. Mike Intel Celeron M 1.5 GHz, 1 GB Ram Microsoft Windows XP Professional 5.1.2600, Service Pack 3 VirtualBox 3.0.10 results from cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Celeron(R) M processor 1.50GHz stepping : 8 cpu MHz : 1476.382 cache size : 1024 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss nx up bogomips : 3007.50 -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group athttp://groups.google.com/group/sage-support URL:http://www.sagemath.org -- Dmitrii Pasechnik - DISCLAIMER: Any text following this sentence does not constitute a part of this message, and was added automatically during transmission.- Hide quoted text - - Show quoted text - -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group athttp://groups.google.com/group/sage-support URL:http://www.sagemath.org- Hide quoted text - - Show quoted text - -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: 3D plots fail using VirtualBox Sage 4.21 on old XP computer
I use Sage all the time on a windows XP machine and firefox, and its usually fine (lots of jmol applets can cause memory problems but that's not specific to that setup). But I am using a remote server usually, not virtualbox. I did briefly test virtualbox on that setup and I think it worked OK for applets. I'll double-check on Monday. -Marshall Hampton On Dec 4, 2:29 pm, Bill Page bill.p...@newsynthesis.org wrote: As reported here in other threads it seems that jmol-based 3-graphics fails in FireFox under Windows XP. I am not sure if it is all configurations anv versions or only some but I do have one laptop running Windows XP and the most recent version of FireFox and this still fails with the newest version of Sage on a remote server. If this is the same problem then I am quite sure it is not a virtual machine problem. It was my understanding that someone (the original jmol developer?) was aware of and working on this problem. No? On Fri, Dec 4, 2009 at 11:59 AM, William Stein wst...@gmail.com wrote: Hi, 1) Try reducing the amount of memory allocated to the Sage Virtual machine from 512MB to 384MB. 2) Windows XP (even with only 1GB) is a very important platform for Sage, IMHO. -- William On Fri, Dec 4, 2009 at 10:21 AM, Michael Madison madison.mich...@gmail.com wrote: I am running Sage 4.2.1 on a old 1GB ram XP computer with the VirtualBox binary distribution. In the notebook if I try a 3D graph Firefox shuts down and kills the Sage session. This is the same problem I had with the VirtualBox Sage version 4.2. I have also created a separate Ubuntu 9.1 vm in VirtualBox, where I loaded all the stuff needed to run and compile Sage 4.21. This included the FireFox upgrades needed to run the 3D plots from a notebook. To this vm I added a second network connection which is a VirtualBox host only, just like in the binary VirtualBox Sage 4.21. I compiled Sage 4.21. When I run the notebook sage using the Firefox inside the Ubuntu 9.1 vm everything works great and I get 3D graphs. I can also start a Sage notebook with a IP address and access it from my Windows Firefox. This works fine until I run a 3D graph and then the windows Firefox closes down, just like in the binary distribution for VirtualBox Sage 4.21. However, if I start Firefox in Ubuntu I can get into the same sheet and run the 3D graph with no problem and Sage has not shut down. I suspect the problem with the 3D graph is from the VirtualBox host connection. I don't know if it is just because of my XP system. -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Simplifying arctan2 expressions
I have been using sage to solve for the eigenvectors and eigenvalues of a 4x4 symbolic matrix. It takes a few hours, but sage is able to return the solutions. My dilemma is that the resulting expressions are *HUGE*. If I ask sage to display one of the eigenvalues on screen (not using show()) it takes several minutes and then finally displays a lng, truncated string. As a specific example, the full text file containing the output of the first eigenvalue.real_part() is 8.6 MB. Sooo I am currently working on trying to simplify the resulting expressions in sage to shrink them down to a smaller size. Here is the problem: I notice that my expressions contain *a lot* of terms like this, arctan2(0, long_expression). Now I know that arctan2(0,x) = 0 for all x and so this should simplify my expressions a great deal. However, I cannot get sage to make this simplification for me. Simplify() and simplify_full() do not touch these terms. Does anyone have an idea about how to address this trigonometric simplification? Thanks so much! -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
[sage-support] Re: losing confidence in Sage notebooks
Dima Pasechnik wrote: I referred to an apparent missing feature of exporting notebook cells into a Sage script. One way to do this is to press control-backspace in the bottom cell a bunch of times until all the cells' code is in one cell (that keystroke joins the current cell with the cell above it). Then select all (control-a, for example), copy, and paste into a sage script file. Jason -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
On Dec 4, 2009, at 9:31 PM, Dima Pasechnik wrote: 2009/12/5 Robert Bradshaw rober...@math.washington.edu: On Dec 4, 2009, at 4:53 AM, Dima Pasechnik wrote: just a side remark - IMHO notebooks are not designed for any kind of large-scope project. There's no reason they couldn't be. I meant a project that takes a lot of computing power (CPU/memory/ disk space). I cannot but agree that for designing, e.g. exercises, notebooks are very useful, esp. combined with functionality of sagenb.notebook. Oh, that's what you mean. Still, I don't see any reason the notebook would be worse. I've let things run overnight in the notebook, and @parallel works there too. For the latter, you are much better off with good old scripts. Notebooks certainly have their own pluses, such as more interactivity and ease of collaboration---but not sharing, as was demonstrated here recently. Otherwise I don't really see a point of them. Ease of use, especially for people on Windows. I think they have advantages for sharing as well, as I can share a notebook with someone by posting it on a public (or personal) server, and they can go and get an account and use it without having to even install sage. This could be particularly nice for the classroom setting. I referred to an apparent missing feature of exporting notebook cells into a Sage script. This seems to be locking a user into using worksheets long after it's time to move over to full-blown development with scripts. True, that would be a nice feature. Clicking on the text link is a good start. - Robert -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
On Dec 4, 2009, at 12:20 PM, tmb wrote: -- For single user mode, put all the worksheets in ~/SageMath (that tells me that I'm supposed to look at them) It's standard to make a .foo directory to hold application defaults and data. Making a visible top-level directory is more invasive. Dotfiles contain application defaults, they do not usually contain application data. Application data is usually given on the command line or stored in a toplevel directory. Many IDEs, for example, create toplevel directories in the home directory that contain their projects (this is often configurable in a dotfile). None that I use I guess. There are no visible top-level folders in my home directory that I didn't put there (except the ones there on account creation) and I like it that way. I guess it's a manner of preference, but at lest it's consistent. -- Put all notebook files at the toplevel with descriptive names, like ~/SageMath/fft.sws Where would the descriptive names come from? How would one handle naming conflicts? Worksheet renaming? The usual thing to do is that documents have a short name that is used in the file system and URL, and a title that is contained in the document and stored in the index. Google Sites uses that for example. The short name usually remains fixed (but can be changed if necessary) and is used for things like merging versions etc. The title can be changed pretty freely. The short name is initially suggested based on the document title. With a directory structure like that, things get a lot simpler for me because I actually understand what I can do: -- Make a backup: cp ~/SageMath/*.sws ~/my-sage-backup -- Restore backup: cp ~/my-sage-backup/* ~/SageMath This works already (with ~/.sage/sage_notebook) Oh? Why don't you try it. cp ~/SageMath/*.sws ~/my-sage-backup doesn't back up my Sage worksheets. A recursive directory copy of the entire tree does, but that's not particularly interesting. Why not? (Or you could copy the whole sage_notebook folder.) Even there, I don't know whether I can safely restore it on another instance. Yes, you can, I've done it. -- Restore a few files: cp ~/my-sage-backup/improc*.sws ~/SageMath -- Fix something in the fft worksheet: vi ~/SageMath/fft.sws Note a .sws file may is actually a archive of several relevant files (e.g. images that have been uploaded into DATA). Sorry, I meant fft.txt (replace .sws with .txt in all the examples). One still has the issue of naming conflicts between multiple systems. Not if worksheets use explicit file names (see above). I think it makes more sense to put a repository inside each worksheet. That's not convenient when I have 100 or 200 worksheets. Do you always want to move around all 100-200 worksheets as a unit? This feels be like having a single repository for all the .tex files on your system, or all images, or something like that. I guess I view my worksheets as individual units that I move/send around. It would still be easy to write a script to sync every worksheet (say, for a given user) one at a time. But the idea of putting this data under revision control is a good one--no one's had time to implement it yet. Do you want to help out? Version control and more sensible naming isn't something I need very much in and of itself; if Sage worksheets worked correctly, I'd mostly be happy with it the way it is. And it is not particularly important to me at the individual worksheet level, since most of the problems I'm encountering seem to be happening at the directory tree level (I forgot to mention duplication of worksheets). I'll probably just work some separate Python scripts for copying and merging notebook directories and keep my fingers crossed that the Sage worksheet issues get fixed. I think they will be. 4.1.2 was a major change to the notebook, things should be stabilizing from here on out. Of course, there's lots of room for improvement, and I do appreciate your feedback. - Robert -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
2009/12/5 Robert Bradshaw rober...@math.washington.edu: On Dec 4, 2009, at 9:31 PM, Dima Pasechnik wrote: 2009/12/5 Robert Bradshaw rober...@math.washington.edu: On Dec 4, 2009, at 4:53 AM, Dima Pasechnik wrote: just a side remark - IMHO notebooks are not designed for any kind of large-scope project. There's no reason they couldn't be. I meant a project that takes a lot of computing power (CPU/memory/ disk space). I cannot but agree that for designing, e.g. exercises, notebooks are very useful, esp. combined with functionality of sagenb.notebook. Oh, that's what you mean. Still, I don't see any reason the notebook would be worse. I've let things run overnight in the notebook, and @parallel works there too. Judging from what I see posted here, notebooks are not very well behaved under extreme circumstances like OS crashes (that might be a result of running out of memory, etc). One is tempted to have unsaved cells open, too. Under a heavy load all this does not play well. Also, a nontrivial project has a non-trivial amount of code that one needs to maintain, too. One essentially designs a small library. A special notebook functionality is needed to handle such a case. IMHO it's a bit wasteful to implement such a functionality. After this is done, one has to wait just a bit to see requests for Sage notebooks to be able to work as an e-mail client :-) For the latter, you are much better off with good old scripts. Notebooks certainly have their own pluses, such as more interactivity and ease of collaboration---but not sharing, as was demonstrated here recently. Otherwise I don't really see a point of them. Ease of use, especially for people on Windows. I think they have advantages for sharing as well, as I can share a notebook with someone by posting it on a public (or personal) server, and they can go and get an account and use it without having to even install sage. This could be particularly nice for the classroom setting. I referred to an apparent missing feature of exporting notebook cells into a Sage script. This seems to be locking a user into using worksheets long after it's time to move over to full-blown development with scripts. True, that would be a nice feature. Clicking on the text link is a good start. yes, but still the result of copy/paste into an editor would require tedious editing (removing sage:, , proper Python indentation), that gets out of hand, particularly, if largish data, e.g. relatively big multivariate polynomials, is involved. Would be nice having some kind of export feature. Dmitrii -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
2009/12/5 Jason Grout jason-s...@creativetrax.com: Dima Pasechnik wrote: I referred to an apparent missing feature of exporting notebook cells into a Sage script. One way to do this is to press control-backspace in the bottom cell a bunch of times until all the cells' code is in one cell (that keystroke joins the current cell with the cell above it). Text button does this even quicker. Then select all (control-a, for example), copy, and paste into a sage script file. But this would still need a lot of editing by hand... Dmitrii Jason -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
On Dec 4, 2009, at 10:35 PM, Dima Pasechnik wrote: 2009/12/5 Robert Bradshaw rober...@math.washington.edu: On Dec 4, 2009, at 9:31 PM, Dima Pasechnik wrote: 2009/12/5 Robert Bradshaw rober...@math.washington.edu: On Dec 4, 2009, at 4:53 AM, Dima Pasechnik wrote: just a side remark - IMHO notebooks are not designed for any kind of large-scope project. There's no reason they couldn't be. I meant a project that takes a lot of computing power (CPU/memory/ disk space). I cannot but agree that for designing, e.g. exercises, notebooks are very useful, esp. combined with functionality of sagenb.notebook. Oh, that's what you mean. Still, I don't see any reason the notebook would be worse. I've let things run overnight in the notebook, and @parallel works there too. Judging from what I see posted here, notebooks are not very well behaved under extreme circumstances like OS crashes (that might be a result of running out of memory, etc). Well, not even scripts behave well when the OS goes down. One is tempted to have unsaved cells open, too. Under a heavy load all this does not play well. Also, a nontrivial project has a non-trivial amount of code that one needs to maintain, too. One essentially designs a small library. A special notebook functionality is needed to handle such a case. IMHO it's a bit wasteful to implement such a functionality. After this is done, one has to wait just a bit to see requests for Sage notebooks to be able to work as an e-mail client :-) See the email() command :). For the latter, you are much better off with good old scripts. Notebooks certainly have their own pluses, such as more interactivity and ease of collaboration---but not sharing, as was demonstrated here recently. Otherwise I don't really see a point of them. Ease of use, especially for people on Windows. I think they have advantages for sharing as well, as I can share a notebook with someone by posting it on a public (or personal) server, and they can go and get an account and use it without having to even install sage. This could be particularly nice for the classroom setting. I referred to an apparent missing feature of exporting notebook cells into a Sage script. This seems to be locking a user into using worksheets long after it's time to move over to full-blown development with scripts. True, that would be a nice feature. Clicking on the text link is a good start. yes, but still the result of copy/paste into an editor would require tedious editing (removing sage:, , proper Python indentation), that gets out of hand, particularly, if largish data, e.g. relatively big multivariate polynomials, is involved. Would be nice having some kind of export feature. Sorry, I meant click on the edit button, much easier to edit down to just the code from there. - Robert -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
such a functionality. After this is done, one has to wait just a bit to see requests for Sage notebooks to be able to work as an e-mail client :-) See the email() command :). huh? It does not seem to be possible to locate documentation on this... Certainly not in various indices, and plain Google gives lots fo false hits. -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
Re: [sage-support] Re: losing confidence in Sage notebooks
On 5/12/2009, at 8:29 PM, Dima Pasechnik wrote: such a functionality. After this is done, one has to wait just a bit to see requests for Sage notebooks to be able to work as an e-mail client :-) See the email() command :). huh? It does not seem to be possible to locate documentation on this... Certainly not in various indices, and plain Google gives lots fo false hits. sage: email? Type: function Base Class: type 'function' String Form:function email at 0x10a2336e0 Namespace: Interactive File: /Applications/sage/local/lib/python2.6/site-packages/ sagenb/notebook/sage_email.py Definition: email(to, subject, body='', from_address=None, verbose=True, block=False, kill_on_exit=False) Docstring: Send an email message. INPUT: to -- string; address of recipient subject -- string; subject of the email body -- string (default: ''); body of the email from_address -- string (default: usern...@hostname); address email will appear to be from verbose -- whether to print status information when the email is sent block-- bool (default: False); if True this function doesn't return until the email is actually sent. if False, the email gets sent in a background thread. kill_on_exit -- bool (default: False): if True, guarantee that the sending mail subprocess is killed when you exit sage, even if it failed to send the message. If False, then the subprocess might keep running for a while. This should never be a problem, but might be useful for certain users. EXAMPLES:: sage: email('xxxsageu...@gmail.com', 'The calculation finished!') # not tested Child process ... is sending email to xxxsageu...@gmail.com NOTE: This function does not require configuring an email server or anything else at all, since Sage already includes by default a sophisticated email server (which is part of Twisted). -- http://yomcat.geek.nz -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org