I just bought the scrivener software this morning after using the trial version for the last couple of weeks; it worked great until late last week, when I imported ~2000+ files (pdfs mostly) into the Research section. Since then it has become super-slow – if I go from, say, Word to Scrivener, it takes anything from 15 seconds to a minute for Scrivener to come up active on my screen.
Has anyone else encountered this problem? Is this a function of having so many imported files linked to Scrivener? Is it an issue with Snowleopard specifically? How can I get back the original speed?
I would like if possible to not have to delink the pdfs from scrivener. I’m writing my dissertation, and being able to use the split view was one of the biggest attractions…
Sorry you’re encountering problems. There have been some reports of typing slowing down in some situations on Snow Leopard, which I’ve been working to resolve, but nothing like what you are reporting. Were you using full screen at all? And what happens if you quit Scrivener and restart? Is it still slow then? It is possible that if there is a memory leak somewhere, that when importing lots of documents the memory could have got chewed up. If that’s the case, though, it should clear as soon as you restart Scrivener. (That would still indicate a bug, mind.)
Thanks for responding so soon Keith… I’ll try to answer all your questions, and add anything that I can think of.
I never use full screen, and I’ve restarted Scrivener several times in between, to see if brute force makes it faster, but it hasn’t helped. The first time I tried importing all my files, it froze at ~1600 files and never completed the import, which I then cancelled. I actually deleted the imported materials and reimported them again this morning, hoping it would help, but it hasn’t (except psychologically?).
Btw, anytime I SAVE, it slows down even more, to upto a minute and a half, two minutes. This slows down my shifting to other apps too (Word, Excel, Firefox, Adobe), though I have plenty of memory. (My machine is all of 2 months old – it’s FAST, and has plenty of space.)
I wonder if pdf files are the issue, because my husband’s powerbook has problems handling too many open pdfs (think 20+) at once since he upgraded to snowleopard – it freezes, much like Scrivener has started doing.
Just out of curiosity, how big is your Scrivener project (select it in Finder and press Cmd-I)? A while back we had someone reporting really slow backup to zip times, and it turned out they were working with a Scrivener project that was two or three gigabytes in size.
Yes, you can reference PDFs, as opposed to embedding them in the project. In the Inspector, if you click the button that looks like a little bookmark, you can drag URLs, files, and other Scrivener documents into this list. It just holds a list of the items and loads them in their respective viewers when you double-click on them.
Please ctrl-click on the .scriv file in the binder and select “Show Package Contents”, then look at the files inside in the list view of the Finder. Let me know the file sizes for:
binder.scrivproj
binderStrings.xml
binder.scrivproj is the one that gets saved when you hit save, so that would have to be massive to slow down saving. binderStrings.xml may well be enormous if you have imported 1,600 PDF files, which is a heck of a lot of research for one project.
binder.scrivproj is 768kb, and binderStrings.xml is 118.4; not huge, right?
but the .scriv file is 1.19GB…should that matter? (and it’s dissertation writing time – i may not actually USE all 1,600+ files, but i’m afraid i may need something that i read 4 years ago and not find it!)
the pdf files open fine in preview, and actually in adobe too. adobe just doesn’t seem to like too many open files at once…
thanks a lot for taking the time to rack your brains on this… appreciate the help!
an update: i’ve gone, once more, the brute-force route. i’ve made multiple .scriv projects, with one (extremely slow) master copy, with the 2000+ research articles linked to it, and other little associated .scriv projects that i (have already and will continue to) import into the master project. Little chapter projects, which have a few hundreds of files associated with them, are working more or less okay speed-wise. Master project’s still a major pain.
please let me know if you figure out what’s happening, and how i can get around it… i would still like to use the master project, since having the dissertation in the little pieces freaks me out.
Well your project definitely is on the larger side of things, though in your case it’s entirely due to PDFs. The Binder.scrivproject file is definitely on the bigger side, but the strings file is small, indicating you haven’t written a ton yet. The ~2gb total size could be causing issues if your computer is older. What you may wish to consider is using a separate application for all of these PDFs, something that is more suited to such a large quantity of files. Scrivener is great in the small to medium range (strictly in terms of bulk research like we are talking about here—it’s plenty mighty in terms of the text within large writing projects), and can certainly function in the massive range, but that isn’t where its strengths are. You might want to take a look at something like Yep, which is designed all around PDFs and keeping them organised. Another application that a lot of academics use is something from the DEVONthink line. Either one of these will most likely handle ~2gb of PDFs better, and will keep your Scrivener project lean and mean. One reason for keeping it smaller is that it makes backing the project up frequently a less troublesome task. A lot of people like to backup at least once a day, sometimes more often when highly active with the file. With a project that large, you’ll find yourself running out of space and sitting around waiting for backups to zip.
While it might be less comforting than having everything in one place, having two tools in this particular scenario would probably be your best move. A tool for organisation of research and a tool for writing.
Your binder.scrivproj file is fine, but the binderStrings.xml is obviously rather large at over 700MB and would take some time to save. The thing to remember is that when Scrivener auto-saves, only the binder.scrivproj (and any text files that have been edited) gets saved. When you manually hit cmd-S or go to File > Save, the binderStrings.xml is saved too. This keeps auto-saving much quicker. The binderStrings.xml file gets saved no matter what when you quit, and if the app crashes it can always rebuilt, because it’s not critical like binder.scrivproj.
So, the question is, are you manually saving a lot? If you are, then I would expect you to have to wait a while as that 700+MB file writes to disk (about as long as it would take to copy such a file on disk via the Finder).
I might have read these numbers wrong. I read that the binder.scrivproject file is ~700 kilobytes and that the BinderStrings.xml file is ~120 kilobytes.
If there’s 1,600 PDF files in the project, I doubt the binderStrings.xml file could be as small as 700kb - that’s why I assumed megabytes. vasudha, could you clarify please?
Thanks,
Keith
binder.scrivproj is actually only 768KB, but binderStrings.xml is 118.4MB. Forgot to put the units the last time.
I do save a lot manually (force of habit from using Word, and usually only realize it when Scrivener stops responding…), but from what you said earlier Keith, it seems like it should not slow down a 768KB binder.scrivproject file. Also, I don’t always have a pdf open, but it’s still SLOW. And Amber, my machine is brand new, has 2GB 800Mhz memory and 2.13GB processor speed, and >250GB hard drive space free…from what you said, it seems like the 2GB total size should not pose problems in my case (thanks for the tip about the pdf softwares btw.)
Sooo…if I stop manually saving, the Scrivener-suddenly-slowing-down problem should go away? How come it worked so beautifully the first three weeks that I used it in spite of my manually saving all the time? (I’d like to have the problem stop, but would also like to know why it happened anyway…)
Probably because you hadn’t imported 1,600 PDF files at that point?
Try it out and see what happens. The binder strings file gets saved whenever you manually save, so at 118MB it would take a few seconds to save. Try copying the binder strings file inside the Finder to see how long it takes to copy. Is it about the same amount of time as it takes Scrivener to save? Or is Scrivener really much slower than that?
Hmm, you should really have to make separate projects for each chapter really. On the other hand, 1,600 PDF files is rather extreme, as Scrivener isn’t optimised as a database in the same way that programs such as DevonThink are.
Could you just confirm that this slowdown is definitely solely related to manual save times? If so, then that at least means I know exactly what it is, and in the future I could always add a preference or something for when to save search indexes, to speed things up on large projects.
yup keith, been using different project files all morning, and the slowdown is definitely only when i manually save. sorry about the hassle of the last two days…
btw, i JUST saw the NANOWRIMO offer – wish i’d seen it when i paid for scrivener 2 days ago, since i don’t have any funding this year! sure could have used the additional 20% discount since i’m literally counting pennies …
in any case, thanks for building this software! i’m a little ocd re. writing, and i’ve found scrivener amazingly liberating!
good luck, and i’m sure i’ll be back with more questions soon!
Thanks for the info, that’s useful. I have it on my list for 2.0 to look into way so optimising the saving of the search indexes (binder strings) to disk, so I hope I can find some solution to speed that up too. Drop me a line next year when 2.0 comes out and remind me of this thread and I’ll see what I can do about cheering you up about missing out on the discount this time around.
All the best,
Keith