On Sat, May 15, 2010 at 5:02 PM, Robert Bradshaw <rober...@math.washington.edu> wrote: > On May 15, 2010, at 4:20 PM, Minh Nguyen wrote: > >> Hi folks, >> >> On Sun, May 16, 2010 at 6:21 AM, William Stein <wst...@gmail.com> wrote: >>> >>> Hi, >>> >>> It's about time http://sagemath.org got totally reworked for improved >>> usability and simplicity. I think a good way to start would be a >>> thread in which everybody who has any thoughts about "stuff being hard >>> to find at http://sagemath.org"; simply makes their personal pet peeve >>> list. So respond and do so. >> >> Please test out the Sage website for usability issues and report any >> problems here. Most of the time, it can be difficult for Harald and >> myself to know any problems unless people report them. > > I'm not sure if this is new, or has been this way for a while, but the > "Introduction" on http://sagemath.org/development.html should probably be > viewable without scrolling down. (E.g. the list of special interest groups > mailing lists is good to have on that page, but arguably not the most > important thing to have first). > > Also, the main page seems to have gotten really busy. We used to have 6 nice > big links right in the center of the page, now you have to scroll half way > down (on my screen) to see them. Is the Cantor GUI the preferred GUI for > Sage now? (That's the impression I get from the site.) It seems we have half > a dozen places on the main site to "fan" sage, which could be good but is a > bit much. I remember this happening with the old site, where the new links > kept getting bigger and bolder to be able to stand out against what was > already there, and then the cleanness of the new site was very refreshing. > > Now, I'm not a professional web designer, and maybe people like it this way, > but that's my impression.
I agree. > Personally, I got to the downloads, notebook > (though I usually go here directly), or browse the code. Regarding the download page, that is currently *terrifying*. Let's compare our download page to Ubuntu's: Look at: http://sagemath.org/download.html Look at: http://www.ubuntu.com/GetUbuntu/download With Ubuntu you get one big smart download link, and a drop down menu for choosing a download location, plus a nice clean link to alternatives. With our page, before even seeing a download page, you get (count them) 21 links in a busy navigation bar, then 12 more links in another navigation bar, then another area with 3 more prominent links to mailing lists and RSS, followed by a link to big map of all the mirrors and a plot of their status. After these 37 links, some of which are probably only important to webmasters, we finally get a download link. No, we get no less than 21 download links, many of them labeled with acronyms like "C3SL", and some aren't download links -- one is a link to the wikipedia page about metalinks, which are described as "resumeable" (which is misspelled by the way!) Anyway, after looking past 58 links, the 59th is labeled "Not sure what to download? Follow the download and install guide." Note that none of the links you just looked at have anything to do with *what* to download -- they are links to *where* you might download from. Following the "what do download" link, gives a wiki page that I'm pretty sure explains how to figure out what you want if you were using the website I designed in 2007, which doesn't exist anymore. Right now, many of those choices basically happen automatically via javascript, right? That wiki page (http://wiki.sagemath.org/DownloadAndInstallationGuide) also has the phrase "If no binary version is available for your system, download the source version.", but there isn't a link to the source code for Sage, which can be hard to find. There is eventually a link to a longer guide (another wiki page), followed by a link to our install manual (the one in the doc directory included with sage), which starts with "Assumptions: You have a computer with at least 2 GB of free disk space and the operating system is Linux (32-bit or 64-bit) or OS X (10.5.x)." This is slightly misleading: (1) 2GB isn't enough space to build sage. It might be enough for a binary though. wst...@sage:~/build/sage-4.4.2.alpha0$ du -sch . 2.2G . 2.2G total (2) The 10.5.x after OS X should be replaced by 10.4.x, 10.5.x or 10.6.x. --- Anyway, back to the download page. If you then select one of the download sites based on the mysterious codes -- I selected "UW2", since I'm at UW. (Note, there is no UW1 or UW, just UW2.) You get "This is a mirror of Sage - Open Source Mathematics Software. Here, you can download Sage for your system and platform. Not sure what to download? Then follow the download guide. For more information, visit the Sage website." And you really do have to work out exactly what binary to get. What happened to the Javascript we had before where it just figures out what binary you should get? OK, searching further I find that if I get to the download page from the homepage by clicking the "arrow" then it knows I want OS X. However, if I get to the download page from the hompage there by clicking download in the nav bar at the top right, it doesn't know that it should default to OS X. No matter what, even if I use the download arrow so that the OS is known, then the metalinks page gives me links to *all* binaries for all OS's. Metalinks are surely great and all. However, if I just take standard Chrome on OS X, click on the OS X metalink, the amazing result is that I get this file: sage-4.4.1-OSX-64bit-10.6-i386-Darwin.dmg.metalink.metalink which is worthless in OS X for me: flat:Downloads wstein$ open sage-4.4.1-OSX-64bit-10.6-i386-Darwin.dmg.metalink.metalink No application knows how to open /Users/wstein/Downloads/sage-4.4.1-OSX-64bit-10.6-i386-Darwin.dmg.metalink.metalink. Basically the same thing happens in Firefox. When I personally tell people about Sage, I just point them to http://sage.math.washington.edu/home/wstein/binaries/ since at least that's a comprehensible way to get binaries, and frankly I'm scared that the current binary download system is so confusing it will scare them off, since it scares *me* off. Another issue. When you are on OS X and you click to download, it says: "These binaries are only for OS X 10.4, 10.5, or 10.6. They will not work on OS X 10.3." somewhere down the page. Just delete this -- it isn't right at all. The binaries are for what they are for, as listed in the binary name. If the above message really mattered, it should be at the top of the page. It's a holdover really from when people actually used 10.3 (back in 2006, really). The bottom of the download page has links to get "SageTex LaTeX style — this LaTeX style file allows to embed and process Sage code inside TeX files" -- isn't this the standard sage package "sagetex", which comes with sage already? So why is it linked to from the download page? --- Anyway, Ubuntu wins here for a clear simple download link in comparison to us. -- William > Maybe one of the > "big buttons" (if we keep that) could be development, with stuff like code, > devmap, and conferences as sublinks. If we're going to thin things down > again, one key piece of input is seeing where people go. (It's a fairly high > traffic site and we have analytics; I doubt all links get equal numbers of > hits). > > - Robert > > -- > To post to this group, send an email to sage-devel@googlegroups.com > To unsubscribe from this group, send an email to > sage-devel+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/sage-devel > URL: http://www.sagemath.org > -- William Stein Professor of Mathematics University of Washington http://wstein.org -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
