Wikipedia:Reference desk/Archives/Computing/2013 September 29

Computing desk
< September 28 << Aug | September | Oct >> September 30 >
Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


September 29

edit

A smarter YouTube?

edit

A month ago I had some browser issues on my computer, and suspected Adobe as part of the problem, so I completely uninstalled everything Flash, everything Acrobat, and everything Firefox. I then reinstalled one thing a time, which is Firefox only so far, and have found myself in an interesting state:

  • With Firefox's "internal"(?) PDF viewer, apparently I can read PDF files without Acrobat Reader. Sometimes I have to save the file first, then d-click on it, but I've had no noticeable problems without Reader. Who knew?!
  • Flash is a bigger issue, but I'm getting along without it for now. For every site that has something Flash-embedded, Ffx displays a grey box saying "You need the latest version of Flash" message -- but without downloading video, most pages load a lot faster!
  • YOUTUBE IS DIFFERENT. Although a page will sit there for a few seconds with that grey box, but SOME videos will then download and play anyway.

So (finally!), we get to questions:

  1. Is YT detecting that I don't have Flash, and converting to some other format? Seems unlikely, since not all videos do this.
  2. Are some videos uploaded in multiple formats (or uploaded in one format and stored in multiple formats), and YT is switching somehow?
    • How does it do that?
    • How come that doesn't happen for everything?
  3. What else would be fun to know here?

--DaHorsesMouth (talk) 01:48, 29 September 2013 (UTC)[reply]

Yes, YouTube now has limited support for HTML5 video, which does not rely on Flash. It's quite easy to detect when the Flash player isn't available. YouTube's page about HTML5 video support is here; it mentions that not all videos currently work. When someone uploads a video to YouTube, the site seems to do some pretty extensive processing (sometimes videos can take an hour or more to become publicly available) - some of this seems to be scaling (so if the original is HD, it produces a version for each of the lower-resolution options), some of it making it streamable and skippable, and maybe also making format changes so they can support players with different decode codecs. -- Finlay McWalterTalk 11:06, 29 September 2013 (UTC)[reply]
Incidentally, Firefox (and Chrome) does indeed now have its own PDF viewer, where the PDF document is decoded and rendered into HTML+CSS and then is drawn with the same code that renders web pages. Firefox's own info on that viewer is here, and I'm pretty sure that the actual code that does the hard work is pdf.js. Like every PDF viewer (including Adobe's), this process isn't always perfect or consistent with other viewers. -- Finlay McWalterTalk 11:10, 29 September 2013 (UTC)[reply]
And as you asked how Flash is detected, this StackOverflow question shows some methods. Also, some Firefox users who do have Flash installed, but who are vexed by additional burdens (in load time and in CPU and memory use), and by the additional opportunities for mischief Flash affords disagreeable website, run with Flashblock enabled, so pages load with Flash off and the user must make an opt-in action before the Flash code is actually started. -- Finlay McWalterTalk 11:21, 29 September 2013 (UTC)[reply]
Note that because of the popularity of the iPad and iPhone and limited support in other tabket mobile devices like Android an increasing number of sites are supporting HTML 5 video as an alternative to Flash for videos, at least on their mobile sites. Probably the most common exceptions are those for catchup TV and similar services who require complicated DRM. I expect the vast majority of sites only support H.264 (since that's generally supported by mobile devices) except for those in the FLOS and similar movements like Wikipedia who may support something else likely either [[Theora][ or WebM which was problematic until recently for Firefox on Windows given the lack of support, but support should be there now provided your underlying Windows OS supports it. I believe it's also often possible on Linux and Firefox, no idea about Mac OS X but I would presume so. Presuming you do have H.264 support, the lack of Flash will be less limiting for video than it used to be. Nil Einne (talk) 12:40, 29 September 2013 (UTC)[reply]
  Resolved
 – Thanks so much to both of you!

The state of the art continues to advance. --DaHorsesMouth (talk) 03:09, 2 October 2013 (UTC)[reply]

Non-response of MS Excel Program

edit

I have a PC with Windows 7 OS and MS Office-2000. When I try to open a Excel file,I get the foll. msg." There was a problem sending command to the program " and the Excel page opens. Then the file can be accessed from the File> Open menu.Also in the Page Break view manual adjustment of the page cannot be done. This program used to work OK till recently.120.60.36.52 (talk) 09:07, 29 September 2013 (UTC)[reply]

So, what changed on the PC between when it worked and now ? StuRat (talk) 05:48, 1 October 2013 (UTC)[reply]

2 wvdial and load balancing

edit

Is it possible to run 2 instances of wvdial each for 1 usb modem and use load balancing stuff so I could use both internet? In short, is it possible to make Connectify Dispatch in linux? — Preceding unsigned comment added by 140.0.229.26 (talk) 09:27, 29 September 2013 (UTC)[reply]

  Note: Link to WvDial - 220 of Borg 10:18, 29 September 2013 (UTC)[reply]

Cache Memory in virtual memory environment .

edit

I am unable to get how cache works in virtual memory environment. I tried to read many articles but did not succeed. Also I am unable to understand How Cache memory is related to page size,multilevel Page tables,Page table entries. I think that I am missing something very important. Please help me out. Any references are also welcome. — Preceding unsigned comment added by PRADEEP 24888 (talkcontribs) 10:43, 29 September 2013 (UTC)[reply]

Perhaps surprisingly, I don't believe there's that much interaction between the two. The point of the cpu cache is that it's mostly transparent - that the system would work the same (but slower) if it wasn't there at all. So it's done its work, or has failed to do that, before the memory controller actually tries to access actual memory (whether that's actual RAM or address space that's marked invalid and will have to be paged in before the request can be satisfied). The only substantive interactions between the memory cache and the page table (and TLB) that I can think of are:
  • The page table entry has a "is this memory cached?" bit (on Intel that's the Cache Disable bit); if the cache is disabled for this entry, accesses to this memory space aren't cached at all (this is used for accessing memory-mapped hardware and for some kinds of inter-processor communication)
  • The page table entry has a "is executable" bit, which (perhaps only implicitly) controls whether the instruction or data cache will handle the access yes, that's stretching it a bit
  • When a page is unmapped, the OS's memory manager will have to make sure that memory cache entries associated with that are invalidated. On some architectures this may mean the entire cache will be flushed (Intel's WBINVD/INVD instructions flush the whole thing; I don't know if they've latterly introduced a more granular instruction). See for example the description of flush_cache_all on Linux.
For the sake of sanity having page boundaries aligned to cache lines makes sense. All bets are off on weird NUMA architectures and things with per-core TLBs, but then everything to do with caches is harder in such cases. -- Finlay McWalterTalk 15:13, 29 September 2013 (UTC)[reply]

Dipping my toe into Ubuntu

edit

I downloaded Ubuntu 12.04 and loaded it onto a flash drive so I could boot off of it. Really, I just want a taste of it for a while before committing myself to a full install/dual boot. Internet works fine, there seems to be no problems with Libre Office, etc. but I can't play any media files, having tried an ancient .mpg file, some .mp4/H.264 files, and even some .mp3 files. The player would open, tell me I was missing the proper codec (though it did identify it) and then give me an option to close.

So, besides the stupidity of giving me a media player that can't play anything besides the OGG file that it came with, what's really going on here? If I had installed the OS fully, would it be working better (i.e. downloading codecs it didn't natively support)? Is there a way to fix the media player while it's on the stick? Matt Deres (talk) 14:48, 29 September 2013 (UTC)[reply]

Ubuntu does not ship with codecs that might infringe on software patents or that might run foul of anti-copying laws (cf DeCSS). Installing the support is very easy. Note that you should be running 13.04 not 12.04 (unless you have a really good reason); "stable" in Debian/Ubuntu speak means "changes less" not "crashes less". -- Finlay McWalterTalk 15:19, 29 September 2013 (UTC)[reply]
Thanks for the explanation - and the tip about which version to use. Is there any way to update the version on the stick or would that have to wait until a full install? Matt Deres (talk) 10:51, 30 September 2013 (UTC)[reply]
As you haven't installed (really) yet, I'd personally download the 13.04 media and start afresh from that. You can update, if you really want - first run "software & updates" and in its "updates" tab, make sure the "notify me..." setting is "any new version", not "long-term support versions". Then run the update manager (I think it's still called that in 12.4 - if not, it's "software updater"), have it check for updates, and it should say "a new version of ubuntu is available" and will allow you to start the update process. -- Finlay McWalterTalk 11:51, 30 September 2013 (UTC)[reply]
You don't want to update a live version running off a USB stick. If there is no persistence file on the drive, then all of the changes to disk will actually happen in unpaged RAM. If there is one, it will get eaten up by tons of updates to packages. If you've installed it to disk, then an update is fine, although I would personally reinstall from scratch since you won't have much of your own stuff on the system yet. Katie R (talk) 13:40, 1 October 2013 (UTC)[reply]