Jump to content

The Main Website Is Yielding Blank Pages


Enchanted Shadow

Recommended Posts

I'm having a problem with the main site. Several times a day, I can't call up any cache pages. Doing any sort of search or putting in a URL directly leads to a blank page. I might be able to load a few at first, but then everything seems to lock up, and I get nothing but blank pages. After a few minutes, things *might* start working again, but it doesn't take much for the problem to re-occur.

 

Is anyone else seeing this?

Link to comment
Turn off your automated bots and the problem should go away.

This is annoying. I don't have any bots. I do have 8 hide plus 3 from my immediate family. They are in a Safari bookmark. If tell Safari to open all 11 caches in their own tabs, I only get about half of them.

 

I don't asking for 11 is bad thing, it is a completely reasonable thing.

Link to comment

Same here. I don't have any bots. At the time, I was just quickly loading 5-10 caches in seperate windows.

 

I definitely do not like the current restrictions. It's not like I was trying to download 100 pages in 2 seconds. It was more like - at most - 10 pages in 10 seconds.

 

I suppose I don't mind if TPTB want to keep spiders from ripping their entire site. However, I would like them to relax this restriction so that it doesn't catch power users.

 

Seperate from that, however, does anyone know what the current thresholds are, exactly?

Link to comment

I don't asking for 11 is bad thing, it is a completely reasonable thing.

11 pages at once is like saying

 

"gimme gimme gimme gimme gimme gimme gimme gimme gimme gimme gimme"

 

In one second. See if you can say that in one second flat.

 

Now, take that and multiply it by a thousand people.

 

Although you may not consider it, a bookmark that opens 11 pages is essentially a robot. There is no conceivable way that you can read 11 caches at one time.

 

Yes it does stink that you can't download 11 caches at once. But the reasoning behind it is sound. The site is set up so everyone has a chance to view the cache listings. Overworking the server with automated bookmarks or even power clicking the pages makes it so others have difficulty using the site during peak times.

Link to comment

 

Although you may not consider it, a bookmark that opens 11 pages is essentially a robot. There is no conceivable way that you can read 11 caches at one time.

 

Yes it does stink that you can't download 11 caches at once. But the reasoning behind it is sound. The site is set up so everyone has a chance to view the cache listings. Overworking the server with automated bookmarks or even power clicking the pages makes it so others have difficulty using the site during peak times.

 

 

Jeremy, with all due respect, not everyone reviews caches the same way you do. Just because you might not open a bunch of caches in seperate windows for simultaneous review, doesn't mean that others don't, NOR does it mean that they *shouldn't*.

 

How you browse is your business. And how I browse is mine. How do you think people would respond if Google's policy was "Please don't run multiple searches simultanously. If you're looking for information, sifting through results takes time, and there's no reason and no possible way you could really make use of simultaneous searching."

 

If they tried that, their stock (as well as number of hits) would drop like a rock within 48 hours. And people don't even PAY to use their site!

 

I *pay* for the services of your site. If the way I browse involves opening multiple caches at once, there should not be a problem with that. As I said earlier, it's not like I'm trying to do a high-bandwidth spider of your complete site, so that I can have the entire shebang for offline viewing. Opening 5-10 caches simultaneously is not an unreasonable methedology. I do the same on many other sites, as do a lot of people. If your servers can't keep up with the needs of your users, than fixing that is on your side of the net. It's not right to blame the users for it.

Link to comment

Playing the "I'm not getting what I paid for" card seems uncalled for, but I've long found this feature somewhat annoying - mostly because the failure case is hard to recognize. Once you've made the naughty list, there's no real indication what's happened, why it happened, or when it will quit happening. It's pretty easy to trigger this even without using any of the "open all links in tabs" kinds of tools.

 

I do know of one other site that does it, and I like their approach better. Once you've triggered the throttler, there is some delay imposed upon page delivery. I don't know the precise shape of the curve, but let's say that it starts at 1/100 of a second. Every page requested while you're on the naughty list doubles that delay. So the first request gets serviced quickly and you can start reading it while the others are queued. It doesn't slap the server around as it becomes kind of self-governing; it just plain won't hand pages out quickly, The site remains responsible to interactive (non-bursty) users. With the governor I described, that tenth page gets a five second penalty (ten if the timer is reset upon delivery of any page instead of starting just at the request) so you're still delivering useful data to the reader.

 

It's not a big deal - personally, I've just learned to order up PQs to avoid this, which may very well be the behaviour you're wanting to encourage - but I thought I'd offer another approach for consideration.

Link to comment
I have a problem of not seeing gallery images. Could it be related to throttling?

Never mind - fixed!

 

(in another forum about image troubles): I was having the same problem - this thread got me thinking. I see that I had:

 

"Load Images - For the originating website only" as a preferences:content setting on my Firefox. If the images are hosted on Groundspeak.com... Hmmm... Add to excepted sites and fixed! :rolleyes:

 

I was having trouble with the gallery images as well - no longer!

Link to comment
I do know of one other site that does it, and I like their approach better. Once you've triggered the throttler, there is some delay imposed upon page delivery. I don't know the precise shape of the curve, but let's say that it starts at 1/100 of a second. Every page requested while you're on the naughty list doubles that delay. So the first request gets serviced quickly and you can start reading it while the others are queued...

That's a sound system, as you describe. User submits all requests, they are served with successive delays. Consuming the information organically should not be hampered, may not even be noticed...

 

I have run across another throttling mode, seemed fair but is more "in your face:"

 

Users who don't pay (free access) get low-res with feature restrictions.

 

Folks who buy a membership get high-res, full features, but agree to a bandwidth limitation, expressed as bits/day.

 

"Power Users" upgrade their membership (a higher fee), and the bandwidth is increased substantially.

 

This way, if someone wants something, then they pay (accordingly) for it. Otherwise, behaviors adjust. BabelHeads recommendation is more elegant, but just my 2c.

Link to comment

I'd have to agree that it isn't my problem. I am also a paying user. I don't want to click 11 times when I have a perfectly good computer that will do it for me.

 

And yes, I can't read them all, but I do want to look at them all occasionally. Having the delay builtin like a previous poster suggested is a great idea.

 

Instead of devoting effort to making less pages served, perhaps time could be spent on making more pages served?

 

And this would certainly help with the weekend slowdowns also!

 

Paul

Link to comment

 

Playing the "I'm not getting what I paid for" card seems uncalled for, but I've long found this feature somewhat annoying - mostly because the failure case is hard to recognize.  Once you've made the naughty list, there's no real indication what's happened, why it happened, or when it will quit happening.  It's pretty easy to trigger this even without using any of the "open all links in tabs" kinds of tools. 

 

I do know of one other site that does it, and I like their approach better.  Once you've triggered the throttler, there is some delay imposed upon page delivery.  I don't know the precise shape of the curve, but let's say that it starts at 1/100 of a second.  Every page requested while you're on the naughty list doubles that delay.  So the first request gets serviced quickly and you can start reading it while the others are queued.  It doesn't slap the server around as it becomes kind of self-governing; it just plain won't hand pages out quickly,  The site remains responsible to interactive (non-bursty) users.  With the governor I described, that tenth page gets a five second penalty (ten if the timer is reset upon delivery of any page instead of starting just at the request) so you're still delivering useful data to the reader. 

 

It's not a big deal - personally, I've just learned to order up PQs to avoid this, which may very well be the behaviour you're wanting to encourage - but I thought I'd offer another approach for consideration.

 

 

In response to your first paragraph, I only played that card after Jeremy effectively categorized power users as unreasonably greedy. In response to that, pointing out what I did was more than reasonable, in my opinion.

 

In response to your second paragraph, it still doesn't tell you that you're being throttled, and aside from that, it does not address the issue that opening up multiple pages simultaneously is common browsing behavior, and does not deserve to be put into the same category as a robot trying to rip the entire site as fast as humanly possible.

 

In response to your third paragraph, PQ's do not solve the issue for me. I *do* use PQs for my initial stages of searching. But when it's time to finetune that list, I open them up in live browser pages. This lets me see the actual cache page in its most recent form, it lets me see all logs, and it gives me the most up to date information as to the most recent finds. PQs cannot properly handle any of those things. Once I have the pages open, I flip between them and close out the ones I'm not interested in, until I'm left with the final set - which are in the perfect position to be printed.

 

I think I have a nice and efficient way of doing things, as far as my needs and desires go. I'm not doing anything unreasonable, and I take offense at the management's characterization of being unreasonably greedy.

Link to comment

It seems to me that once you're on the servers bad side, it takes less and less to trip the overflow. Eventually, it does go away.

 

Premium members should really get an increase in this. I have NEVER seen a site as strict as geocaching.com as far as bandwidth limitations go.

 

People who are getting this and don't see a reason - remove any prefetching extension you might have like FasterFox.

Link to comment
opening up multiple pages simultaneously is common browsing behavior, and does not deserve to be put into the same category as a robot trying to rip the entire site as fast as humanly possible.

The problem is, on the server side of things, it's impossible to tell the difference between a browser grabbing 10 pages simultaneously, and a robot grabbing 10 pages simultaneously. Thus, they both get throttled. I'm sorry, but I remember the days of endless timeouts and the site being almost completely unusable to everyone. Then this throttling was implemented and suddenly the site was a pleasure to use again. So I completely support the corrent throttling mechanism.

Link to comment
People who are getting this and don't see a reason - remove any prefetching extension you might have like FasterFox.

I don't know about any other prefetching extension, but FasterFox is smart enough not to cause a problem here.

 

From the FasterFox FAQ:

Can prefetching "mess" things up?

Since prefetching is basically the same as clicking on a link, and clicking on a dynamic link can perform some action such as "logging you out" or "emptying your cart", only static content is prefetched by Fasterfox.

...

Fasterfox further limits prefetching such that only files with the extension .gif, .htm, .html, .jpeg, .jpg, .pdf, .png, .text, .txt, and .xml are prefetched.

Since virtually all of the content on geocaching.com is NOT static (i.e. has a ? and parameters in the URL), and most pages here are .aspx, FasterFox will not prefetch any pages from this site.

Link to comment

 

The problem is, on the server side of things, it's impossible to tell the difference between a browser grabbing 10 pages simultaneously, and a robot grabbing 10 pages simultaneously. Thus, they both get throttled. I'm sorry, but I remember the days of endless timeouts and the site being almost completely unusable to everyone. Then this throttling was implemented and suddenly the site was a pleasure to use again. So I completely support the corrent throttling mechanism.

 

 

I'm loading 10 pages, but waiting 2 seconds between each, and I'm *still* having problems. I think it's easy to tell the difference between 1 page every 2 seconds and a request for 10 pages that all arrive within the span of 1 second.

Link to comment

 

Get real. Do you really think they are got to tell the robot authors exactly how to write their programs to avoid the throttling?

 

 

Get real yourself. The idea isn't to prevent robots, because there IS no way to prevent robots. If you're going to do anything, than the idea is to throttle bandwidth so that it's not eaten up completely by robots, leaving nothing left for human users.

 

But ultimately, you do have to cater to the human users. And right now, some of those human users WHO ARE PAYING FOR THIS DAMNED SERVICE are the ones being throttled. I don't think it's unreasonable to lay out the exact tarpitting thresholds on the table, so that we can look at them in detail, instead of speaking about them generically. In addition, it **might** provide a way for those of us having problems to find a way to use the site the way we want to, without triggering this trap.

Link to comment

Flame away at me for saying it, but......

 

Once again, I notice that all the so-called power users that claim its such a hardship not being able to open 10 cache pages at once, only average a handful of finds a month, or less.

 

All the people that don't seem to have a problem getting thottled have find counts in the thousands.

 

Why is it that the people who can find 50-100 caches in one day (over 200 for at least one person in this thread) don't have an issue with throttling, and yet the people who find 25 caches a year NEED to read 20 caches pages a second?

Link to comment

 

Flame away at me for saying it, but......

 

Once again, I notice that all the so-called power users that claim its such a hardship not being able to open 10 cache pages at once, only average a handful of finds a month, or less.

 

All the people that don't seem to have a problem getting thottled have find counts in the thousands.

 

Why is it that the people who can find 50-100 caches in one day (over 200 for at least one person in this thread) don't have an issue with throttling, and yet the people who find 25 caches a year NEED to read 20 caches pages a second?

 

 

Boy, some people really don't have the ability to wrap their heads around the fact that just because someone does something differently than you, doesn't make it wrong.

 

Let me give you an example - just a single example - of how this might work out, since you don't seem to understand the general concept.

 

Possible Example: Not everyone has the free time on their hands to spend 8 hours a day caching. Some people can only manage to do it once every few weeks. For those people, it is possible that - given how few caches they can actually make time for - they want to maximize the QUALITY of those few caches they hit, and so they try to narrow the list down significantly, in order to weed away the ones that are likely to be dissappointing to them (such as doing 20 lamppost magnetic micros in a single afternoon). As such, they need a large pool to choose from, and they need to carefully do their research in paring down the list - as opposed to people who have the time to find 1000 micros per month, and don't need or care to discriminate at all.

 

If you can't simply accept that different people might do things differently, than does this example help you to understand ONE possible way how the phenomemon you describe *might* have come about?

Link to comment
want to maximize the QUALITY of those few caches they hit, and so they try to narrow the list down significantly, in order to weed away the ones that are likely to be dissappointing to them

Do like the rest of us do. Run a few Pocket Queries, and use Watcher or GSAK to do your filtering off-line. It's actually easier this way, because it's faster and you can do full-text keyword searches ;)

 

If you can't simply accept that different people might do things differently,

This isn't Apple. It's obvious that TPTB don't want you to do things differently, here :P Instead of flaming the existing system, why not find a way to work with it? :(

Edited by Lil Devil
Link to comment
Guest
This topic is now closed to further replies.
×
×
  • Create New...