[Date Index] [Thread Index] [Author Index]
Re: Text search within a documentation page?
On Jul 26, 2:24 am, AES <sieg... at stanford.edu> wrote: > In article <g6c9dg$c1... at smc.vnet.net>, David Reiss <dbre... at gmail.com> > wrote: > > > No, AES is right in some cases for the Documentation Center pages. > > Some Documentation Center pages are peculiar in some ways in terms of > > using Mathematica's search capabilities. > > > For example. Go to the main Documentation Center page. Once there= , > > click on the page to be sure that you are in it. Now bring up the > > find and replace dialog and, for example, type in "Overview" (without > > the quotes) in the Find field and click on the Next button. Nothing > > happens (and a warning beep is generated) even though the word > > Overview does appear on the page. You can use the Web version of the "Documentation Center" and it does not have this problem. > > > This sort of thing does not happen for all pages by any means, but > > there are some peculiarities to be found. Perhaps it has something t= o > > do with the style sheets that is used, or perhaps it is a gremlin... > > Thanks for backup on this. > > Another peculiarity is that in the "Search Result" pages (which > sometimes -- often? -- open when you enter a term or select a symbol > name in a Mathematica notebook and click on Help), there are cases where > a search within that Search Result page for, e.g. "Plot" (with Ignore > Case checked) will cycle through all the occurrences of plot or Plot in > the tiny-type text of the links on the page -- but never touch the Plot > string in the titles of the links on the page, e.g. Plot3D. > Online version doesn't have this problem either. But maybe neither does the local version I just discovered. It seems that if you use the "Previous | Next" links to go to another page of the found items, the "Find" window's Next button appears not to work, but if you click on the "Previous" button, the strings are found in reverse. So it would appear that when you use the "Previous | Next" feature, the position of the start of the search is at the bottom of the page, and even though the "Wrap around" checkbox is checked it doesn't wrap around. Sure looks like a bug to me?? > The real question, it seems to me, is, what good _at all_ are these > online Search Pages? If you search on a term, you frequently get Searc= h > Result pages that contain 100s to 1000s of links, extending over many > 10s of further screenfuls or pages. > > How can one proceed further at that point? Is there any way to drill > down further into these very long Search Result lists _efficiently_? I= f > they were structured as one immensely long page, one could at least both > rapidly page down using the Page Down key, to scan all the entries, or > more efficiently jump down to relevant entries using the Find command -- > but I don't find any way to do either of these. > > Notes to WRI's documentation people (if there are any . . .jab, jab): > > 1) Safari has a beautiful search function where, if you've opened an > immensely long web page, you type a search term into the search field at > the top and it immediately shows you the total number of hits in the > entire document, with little forward and back arrowheads on each side of > the number. Clicking either of these instantly jumps you to the next o= r > previous occurrence of the search string in the document (centered in > the document window, of course, which is where any online search of this > type ought to take you). > > Gorgeous technique to use; best one I'm aware of. I agree 100% with this, and on top of that Safari highlights all the occurrences of the searched string on the page so they are very easy to visually pick out. Also if you use the web version of the Documentation Center, this is completely doable right now. > > 2) As someone who does a fair amount of searching in documentation, I > make every effort to capture on my own HD PDF files of the full manuals > or books about software I use (or other technologies or subjects I'm > studying; and then often find that the most efficient way to search for > information on some topic is just to open one of these documents and do > a pure linear full text search in Acrobat. > > This is very often much better than going to the index of the document. = > Indexes can be helpful on paper, much less helpful on screen: too much > awkward jumping back and forth between index and main text. And even > with good intentions, doing a good job of indexing is hard. With the > linear text search you may have to divert your attention to something > else for a few moments, while Acrobat grinds its way through a 500-page > manual -- but when you get a "hit", you're very often "there" (at the > info you want), or have seredipitously learned something useful related > to your search. Sometimes I think that I would love to have a PDF version of the Documentation Center available -- would think is this something that Wolfram could generate?, but know it would be immense, so maybe this is not practical. But the ability to insert my own comments and put in my own bookmarks, and all the other things that you do with PDF files in Acrobat would make this something I would like to have available to find out.