@GustavoBandeira I'm not really sure what you're asking about there? If you're running a custom distributed-memory build of Mathematica on a new publicly-undisclosed supercomputer such that you can reasonably expect to be able to build a list 100TB in size, I think you need to contact WRI or your HPC partner about this...
If on the other hand you had in mind to create a lazy iterator, you'll probably find this superb question/answer of use.
@R.M @J.M. @mr.wizard As you have noticed user Subbu is back again, without having improved anything in his behavior. I propose we suspend him again, this time for a longer period of, let's say, 1 month. Agree?
@J.M. I have some troubles with editing the mod message. The markdown edit box is shifted right over the page smack down in the sidebar and I can see only half of it. Could you try it?
@j.m I saw it; good. I downgraded my graphics driver today, back to the old Dell approved one. My PC got slower and slower and I saw hardware interrupts eating up 10-15% of the CPU time, so I thought the new nVidia driver might have something to do with it. Perhaps I should restart.
@J.M. I knew "peace of mind", but not "piece of your mind". Learned something new. Anyway, I installed the newest driver a few weeks ago (just to see if it resolved my ongoing transparent Texture issues), but I can't be sure that it is causing my slowdowns.
By the way, I have a removable external 1tb hdd, ntfs, w7x64, balbal, and I'd like to be able to browse it and search for files, etc, when it's not available. Ideas?
@J.M. trying Brent now... it's a function involving couple of matrix inverses, so I'm trying to aid it as much as it can to only look in the right place to reduce wasteful computations. But yes, it boils down to something 1D and odd multiplicity
If it wasn't for the update, I did a symbolic link clone of every file, reproducing the folder structure (actually, a software did it). But that's not satisfactory
But I found a desktop indexing software such as Everything, that indexes everything ntfs, very fast, keeps itself updated. But doesn't show the "·$&%$ tree structure
W7 libraries should cache the stuff of removable drives, gr
@R.M Well, Richard Brent came up with another method for multidimensional problems, but you still have the responsibility of providing useful brackets...
@R.M I pronounce it with an F (probably because of where I was born), but most people pronounce it with a softer v sound (which differs from the soft 'b' in many languages)
The above website shows you the distribution of my surname. In 2007 there were 71,065 of us in the Netherlands
@r.m. BTW did you see my mma.SE 'attachment' update? I'm getting farther and farther from the original question, but I found it fascinating to see how much information you could put in such a small gif.
@Verde That file was a 1-to-1 translation of javascript to mma that I did a couple of years ago. It's still faster than the built-in method although not optimized at all.
@Verde :5926532 It's far from ideal, and has a long list of disadvantages, but it's better than a throw-away file upload site. Advantages: it's quick, notebook opens in a flash without any intermediate logging-in, searching, downloading, opening.
@SjoerdC.deVries The only drawback I see is that image storage could be compressed in the future, loosing fidelity and probably turning the thing unreadable
converting GIFs to JPEGs would not be very useful. JPEGs are good for photos but bad for diagrams and graphs, which is the format most of the content takes here
@R.M. That's nice. I agree with a gist solution, or any other type of database (if people feel more comfortable with having a control over it). Then you could say things like "have a look at the code I posted at the newsgroup (codeID: 213124)"
@P.Fonseca I guess the advantage of using the SE image uploader like Sjoerd did is that the decision to delete or yank the source is out of the user's hands once submitted. With gists, you can always delete if you have control (or email if anonymous)
Leonid was working on some kind of add-in functionality for code storage. I wonder at what point he is, and if it can be adapted for this kind thing. We need his 2 cents on the subject.
If the problem is the gists way of working, just go with a simple database table, managed by our StackExchange account moderators. The advantage is that we can also use it for posting code of other sites (like the mathgroup), we can do search on the database, by author, etc. And this is exactly the point of Leonid project. A way of managing posted code.
@R.M That's a pity, it costs money. I was just looking into their API. It uses the POST method, which is still experimental in MMA but probably would work.
@P.Fonseca Speaking of that... I do run the stackmma github account. We've used it to host some sample files from questions and a notebook for the blog. The long term goal is to be able to easily host notebooks if and when people want it