Interviewed by CNN on 02/05/2012 13:03, hawker told the world: > The way download accelerators have always worked, AFAIK is the the > assumption that a accelerator company (it used to be your ISP) has a > faster pipe. It downloads it, re compresses it and sends it to you that > way, then your end decompresses. This step takes a host server. This is > why they mostly don't work per the test on the net (IE files tend to > already be as compressed as they can, and we now all have fat pipes). > They are a leftover from dial up days that worked with non compressed files. > > If you have different information, and modern download accelerators > don't work like the old ones, than I would love to be enlightened.
The way most download accelerators work is by opening several (usually four, but this varies by product and it's often adjustable by the user) simultaneous downloads of the same file: one starting at the beginning of the file, one at the 25% mark, one at the 50% mark and one at the 75% mark, for instance. The theory being that if there's a bottleneck somewhere, you get a larger share of the bottleneck. Like, if there's 1000 people trying to download from this server, you get 4 parts in 1003 instead of 1 part in 1000. How they do this? They leverage a feature of most modern FTP and HTTP servers (and clients, for that matter): the ability to resume an interrupted download from the interruption point instead of starting all over again. Essentially, the download manager lies to the server, saying it already has the previous chunks. A few servers (not many, it's rare I run into this) have filters to block this behavior. But in most cases, the only thing that happens is that you fall back to plain old single-threaded downloading, that is, you don't lose anything, you just don't gain anything either. (Nirsoft.net, for instance, does claim to block aggressive download managers, but their files are so small that most download managers don't bother splitting the download) I have been using download managers for years, yeah, even in my dial-up days -- my GetRight subscription dates from last millenium, and I used it in "shareware" mode for a couple years before registering. For a long time I depended on FlashGot to integrate GetRight into Mozilla/Seamonkey/Firefox. I still keep it as an option, but DownThemAll is so much more convenient and powerful that it's a rare event indeed that I dust off GetRight nowadays. The real power in DtA is its ability to "scrape" (analyze) the page you are visiting and find all the download links, categorize them and offer them in a convenient interface to select the downloads. You can even generate your own download links with a simple syntax. For instance, if you tell it to download file[0001:1000].jpg it will attempt to download every image named file0001.jpg, file0002.jpg and so forth, until file1000.jpg. By the way, every time someone suggests that I "ought to" move to Chrome because it's the Big New Fad, I point out the lack of anything resembling DtA in the Chrome platform. And I HAVE looked; the so-called "replacements" do only queue-management and sometimes download acceleration, which are, like, 1% of what DtA does. Apparently the Chrome extension API is just too basic for more than that. -- MCBastos This message has been protected with the 2ROT13 algorithm. Unauthorized use will be prosecuted under the DMCA. -=-=- ... Sent from my Multivac. * Added by TagZilla 0.7a1 running on Seamonkey 2.9 * Get it at http://xsidebar.mozdev.org/modifiedmailnews.html#tagzilla _______________________________________________ support-seamonkey mailing list [email protected] https://lists.mozilla.org/listinfo/support-seamonkey
