Re: Convert Byte() to string
- From: "mayayana" <mayaXXyana@xxxxxxxxx>
- Date: Wed, 6 Aug 2008 20:47:38 -0400
Thanks for this info on the "helpers".helper
Do you know of any way for the end user to learn about how badly the
is hurting to web site being accessed?
I've never used a "helper" before myself, so
my only experience is with my own site. I like
to read the server "raw" logs daily, to see
where people linked from, find out when I
have bad links, etc. That's how I started seeing
the problem of download helpers. I don't know how
much the end user can know. I send a 403 to
helpers on my site, but none of the helper programs
seems to be designed to recognize a 403 response.
So there's no way for me to tell people why they're
being blocked.
As I understand it, helper programs originated
with dial-up, as a way to make sure that a download
didn't get dropped partway through. They typically
work by opening numerous connections to the server,
each requesting only part of the download file. That's
a somewhat questionable logic with highspeed, except
when one is downloading something like a 700MB CD
ISO. And it also, often, doesn't work anyway because
many servers won't allow multiple connections from the
same IP. But beyond the issue of their real usefulness, from
reading my logs I find that helpers usually download several
times the bytage of a file. On my own site I have few
downloads over 100 KB, yet I often see people on highspeed
connections downloading a 50KB file in 5 or 6 pieces.
By the time it's done they've downloaded maybe 150 KB
total. Or sometimes they just download the whole file numerous
times within a few seconds. Often they send the server a
browser userAgent on the first download and then download
several more times with a custom helper userAgent string.
Ironically, people don't notice because with highspeed
it only takes a few seconds either way. I often wonder
what people experience when they've downloaded successfully,
say, twice, and were blocked four times, all within 8 seconds.
Did they get the file or do they see it as blocked? :)
The problem for webmasters is 1) excess traffic that's
unnecessary and 2) difficulty counting actual downloads.
I can't identify all helpers with certainty because they
vary in how they identify themselves to the server. The popular
ones like FDM and Download Accelerator seem to be minor
nuisances, downloading 2-6 copies instead of 1. I have reason
to believe that GetRight is by far the worst, calling the server
over and over again, often hundreds of times. (GetRight shows
a dummy userAgent, so I had to identify it based on several
bits of circumstantial evidence.) I don't know whether it's just
badly designed or whether maybe it's outdated.
The behavior makes no sense. But in general all of the helpers
I've seen seem to be poorly designed in that they download
multiples, clearly not keeping track of what they're doing.
(And in most cases these are multiple downloads with a 200
code, not a 206. In other words, they're not downloading
pieces. They're downloading the whole thing redundantly.)
Another annoyance is the programs that download one of
everything linked from a page. I suspect the people who do
that are probably also the ones who never get around to
looking at what they've downloaded. They're too busy hoarding
more files. But those programs (scrapers?) at least seem to
work properly. I seldom see a "scraper" download more than it
needs to.
.
- References:
- Convert Byte() to string
- From: XP
- Re: Convert Byte() to string
- From: mayayana
- Re: Convert Byte() to string
- From: XP
- Re: Convert Byte() to string
- From: mayayana
- Re: Convert Byte() to string
- From: Paul Randall
- Convert Byte() to string
- Prev by Date: Re: Open 'Disk Properties' via rundll32?
- Next by Date: Re: Variable within Object Reference Line
- Previous by thread: Re: Convert Byte() to string
- Next by thread: Re: Convert Byte() to string
- Index(es):
Relevant Pages
|
