New Iteration Of Gpx

1.2k · October 24, 2005

Ok.. I'm starting to gather information for the next iteration of the GPX file.

** editing to update list **

Images that are either excluded, included via link or included via inline.
Travelbug Types.
Cache Attributes (including images either via link or inline)

and....

Tell me what you would like to add/change/remove from the current GPX specification.

Thanks

2.7k · October 24, 2005

Dreaming here, but parking coordinates would be nice, if the user has chosen to add them.

Thanks for asking! Is there an account somewhere with which I could bribe thee?

9.2k · October 24, 2005

Another vote for parking coordinates being a part of the Nearby Parking attribute.

9.2k · October 24, 2005

Also, the date of the last status change. The GPX file contains the status of the cache (active, disabled, archived), but there's currently no way to determine when it changed to that state.

And not a change to the GPX file per se, but it would be nice to have some flexability in the number of logs retrieved. Being able to select a range, from 0 to 10 perhaps, would be nice.

5.3k · October 24, 2005

Oh, and this will sound really picky, but please don't use "GPX", "Pocket Queries", and "Groundspeak's namespace as used by PQs" interchangeably. The geocaching crowd has a tendency to do that and it's a real pain to deprogram someone that's convinced that "GPX" means "Pocket Query".

1.1k · October 24, 2005

How about a time in the found log area other than the same time for all of them?

Most people (most) log in the order in which they found things even if it's days later. At least then when people use 3rd party software the caches would show up in the correct order.

Please refer to thread for a better explanation!

9.4k · October 24, 2005

Some of these requests aren't related to the Groundspeak's GPX namespace. The request is for extending the namespace to include existing data, such as attributes, not extending existing features so they can be included in the GPX Groundspeak namespace. Thanks.

Edited October 24, 2005 by Jeremy

October 24, 2005

... parking coordinates would be nice, if the user has chosen to add them.

Not to get too off topic but wouldn't that first require that the cache submit page actually have a place to specifically put parking coords before they could end up in the GPX file?

edit: Jeremy got that in ahead of me.

Edited October 24, 2005 by mini cacher

2.7k · October 24, 2005

Evidently I have no idea of what I am talking about. Nevermind.

8k · October 24, 2005

Some of these requests aren't related to the Grounspeak's GPX namespace. The request is for extending the namespace to include existing data, such as attributes, not extending existing features so they can be included in the GPX Groundspeak namespace. Thanks.

Looks like the following data are not currently in the GPX file:

Cache attributes
Bookmark lists that contain this cache
Images uploaded to the cache page
Images uploaded to a log
Subscriber-Only indication
Travel Bug types
Total number of logged visits

2.7k · October 24, 2005

What/where is the "current GPX specificaiton"?

Does it apply to benchmarks? If so, I might have some suggestions of things to add.

6.4k · October 24, 2005

Images that are either excluded, included via link or included via inline.
and....

Tell me what you would like to add/change/remove from the current GPX specification.

Thanks

If you are going to include inline images, it might be a good idea to have some sort of limits (besides having the default to no images).

I can think of quite a few caches with dozens or even hundreds of images; especially older caches, events, and any cache Planet finds. Times that by 500 caches in a PQ and I can imagine email attachment sizes getting out of hand real quick, especially since jpgs don't really compress.

3.8k · October 24, 2005

If it is possible I would like to have a check mark on the cache page that allows just photos from that cache included in a pocket query. I don't want all the photos but some caches require the photo to find the cache. Those I want.

7.2k · October 24, 2005

You won't need to add an addition field in the Groundspeak namespace for parking coords.

Add a new way to enter parking coords--as a new log type, a field in the description, or break out all of the coords for the ultimate in flexibility--and simply add an additional <wpt > entry. Simple.

You can do this with any other coords a cache owner wants to provide.

9.4k · October 24, 2005

Add a new way to enter parking coords--as a new log type, a field in the description, or break out all of the coords for the ultimate in flexibility--and simply add an additional <wpt > entry. Simple.

Pretty much my thinking, but I'm trying to keep it on topic to the current site features. So regardless of the new tags in the namespace we can still add parking coordinates when that feature becomes available.

7.2k · October 24, 2005

On the subject of photos in the namespace. You don't really have to include the actual photo, just the URL (possibly a shortened version to save space) and let the GPX processors handle the additional information. New versions of GPXspinner and GSAK could produce thumbnailed HTML pages and when you Plucker/iSilo your local pages the actual images will be grabbed.

Of course, there are other ways to handle the caching of photos, but the point is you don't have to send down the actual photos with the PQ and still get what you want.

Personally, I'd include only the cache's photo with the description and leave the log's photos with the logs--kind of like the coords entered with the logs.

October 24, 2005

On the subject of photos in the namespace. You don't really have to include the actual photo, just the URL (possibly a shortened version to save space) and let the GPX processors handle the additional information.

Maybe the most desirable for the user to keep the email smaller, but (as pointed out elsewhere) doesn't that then require the third party app to "scrape" things from the web site? Or does the direct accessing of an image by a third party app not violate the site usage policy? Just thinking.

9.4k · October 24, 2005

A "Include images" checkbox on the Pocket Query edit page could accomplish both ways of thinking.

5.1k · October 25, 2005

A "Include images" checkbox on the Pocket Query edit page could accomplish both ways of thinking.

if this really happend you might be god to someone.

happy hunting from hamburg.

7.2k · October 25, 2005

...doesn't that then require the third party app to "scrape" things from the web site?

That's not scraping. It's hotlinking. I've not seen anything in the TOU/TOS. Additionally, Plucker/iSilo hotlinks images from where ever they are if the software is set up to do it. It doesn't care if the image in on gc.com or hosted elsewhere.

The only difference would be it would be hotlinking to additional images beyond the ones that have a IMG tag in the description.

October 25, 2005

...doesn't that then require the third party app to "scrape" things from the web site?

That's not scraping. It's hotlinking. I've not seen anything in the TOU/TOS.

Thats what I figured. So as long as that doesn't trigger the IP ban since it would posibly be a large volume of hits in rapid sequence as the GPX file is processed.

5.1k · October 25, 2005

… That's not scraping. It's hotlinking. …

that's splitting hairs. and it’s counterproductive to what a lot of cachers want: the full description of caches. and that means: include the cache belonging images. it’s just a small step to TPTB to satisfy their PMs.

happy hunting - the old fashioned way (trads only, exceptions may appear)

edit: syntax

Edited October 25, 2005 by HHL

October 25, 2005

What/where is the "current GPX specificaiton"?

Does it apply to benchmarks? If so, I might have some suggestions of things to add.

I agree can anything be done with gpx files for benchmarks?

October 25, 2005

What/where is the "current GPX specificaiton"?

Does it apply to benchmarks? If so, I might have some suggestions of things to add.

I agree can anything be done with gpx files for benchmarks?

I will third that one

7.2k · October 25, 2005

that's splitting hairs. and it’s counterproductive to what a lot of cachers want: the full description of caches. and that means: include the cache belonging images.

The two things aren't even close so there is no hair to split. Like I said, anyone who creates HTML with GPXspinner or GSAK is already doing it if the image is referenced in the description. A separate tag which references each image attached will give you each image without much bloat.

I suspect if images were included with the GPX file that is mailed to you, it would be huge. It would eat up bandwidth unnecessarily. It might even make the mail servers choke. To counter that, the number of caches included in a query would then have to be made smaller.

No, thanks. I'll continue to get the images just like I'm getting them now.

7.2k · October 25, 2005

What/where is the "current GPX specificaiton"?

Does it apply to benchmarks? If so, I might have some suggestions of things to add.

I agree can anything be done with gpx files for benchmarks?

I will third that one

I vote against it.

There is already a way to ceate GPX files for benchmarks. All you got to do is ask around.

Go HERE to get started.

Hope this helps.

October 25, 2005

that's splitting hairs. and it’s counterproductive to what a lot of cachers want: the full description of caches. and that means: include the cache belonging images.

The two things aren't even close so there is no hair to split. Like I said, anyone who creates HTML with GPXspinner or GSAK is already doing it if the image is referenced in the description. A separate tag which references each image attached will give you each image without much bloat.

I suspect if images were included with the GPX file that is mailed to you, it would be huge. It would eat up bandwidth unnecessarily. It might even make the mail servers choke. To counter that, the number of caches included in a query would then have to be made smaller.

No, thanks. I'll continue to get the images just like I'm getting them now.

Ok, let's consider the hot linking approach: I have, just as an example, 3000 caches in GSAK, collected with several PQs. Now I want to create an offline database for my palm or PDA. In GSAK, I click "Export to cachemate", GSAK resp. the convertion tool behind it does some crunching and seconds later I have my self-contained database for mobile offline use.

If the images are all hot-linked, the crunching time is the ultimate time when the actual images would have to be retrieved. Maybe hundreds if not thousands of images must be pulled automatically from the servers at that time. Many server hits within a short period of time. To me, this seems to be a good way to get yourself banned, regardless of the naming of it, - unless, certain programs to perform this action are privileged by TPTB, and only those few (specially adapted) programs can and may do the trick without negative consequences (to the user and the servers)...

If you include the images directly with the PQ then the PQ-server can do load balancing and retrieve the images at the times that fit best - and not when some user chooses to click "Export to offline database". I admit that mail size might be an issue. But here, an automatic resize to typical PDA resolutions (e.g. this can be configured at the PQ's setup page) might help to reduce file sizes.

Personally, I would prefer a "once I have the PQ I have everything ready for the offline database" approach, as it makes me less dependent to having continuous access to the servers.

Just my thinking.

HoPri

October 25, 2005

I admit that mail size might be an issue.

Not half! And your approach would see the GC.com servers sending out hundreds, perhaps thousands of images to every user, every day!!!(*) And all because they want the option of grabbing a single spoiler pic if they get stuck! No, I think URLs are the way to go.

Photos don't actually change that often, so sending them out time after time is a complete waste of bandwidth. The fact that the GC.com servers can choose the best time doesn't get over the fact that your scheme would require hundreds, perhaps thousands of times as much server resources in total.

Other programs like SpoilerSync grab their images in a manner which is sympathetic to the GC.com server (lots of pauses, lots of caching, lots of filters so you can only use bandwidth where the description contains the word "SPOILER" etc, etc). I'm sure Clyde can upgrade GSAK to make it equally so. (With bells on, if past experience is anything to go by! :huh: )

But here, an automatic resize to typical PDA resolutions (e.g. this can be configured at the PQ's setup page) might help to reduce file sizes.

It'd need to be configurable - just because you chose a device with a tiny screen, doesn't mean we all did! And anyway some photos need to be zoomed in order to be useful.

My vote is for new tags in Groundspeak:cache and Groundspeak:log, which contain the URL, width, height, filesize and description of both thumbnail and full-size photo.

Cheers,

Ian

(*) well, OK, only those who set up unnecessarily large PQs which run unnecessarily frequently. Which is most people, from what I can tell :unsure:

October 25, 2005

Ok, you have a point. Not all PQs result in offline databases, and you would save retrieving images for those if you do hot-linking. But for those PQs that are turned into offline databases, I assume that the total bandwidth for retrieving the images is the same, irrespective of the the eventual timing. Then one might want to control the timing to make efficient use of the available server power. And isn't offline databasing the main use-case for PQs (i.e. take out on your PDA) ? Therefore, I rekon that the big majority of created PQs are eventually turned into an offline database. But that's just my guess, and I agree. The bandwidth argument depends on the relation of PQs turned into offline databases to the total number of created PQs. If it's low then sending the images with the PQ wastes bandwidth.

Other programs like SpoilerSync grab their images in a manner which is sympathetic to the GC.com server (lots of pauses, lots of caching, lots of filters so you can only use bandwidth where the description contains the word "SPOILER" etc, etc). I'm sure Clyde can upgrade GSAK to make it equally so. (With bells on, if past experience is anything to go by! )

On the other hand and to my understanding, the TOU do not allow the use of programs like spoilersync, regardless how little stress they provide for the servers. Hence, the hot-link option would require that at least some pieces of software are granted the right to automatically retrieve the images upon request.

My vote is for a legal (and convenient) way to get the images at all. I have a preference, but if it's this way or the other does not bother me too much.

Best regards,

Holger

Edited October 25, 2005 by HoPri

7.2k · October 25, 2005

Yes, SpoilerSync does violate the TOU unless permission has been granted. It reads the cache page, extracts the URLs of the photos, and then downloads the photos.

Hotlinking would download the photos directly.

The main advantages of going the hotlinking route is IIRC the images are on a seperate dedicated server, proxy caching (if there is one near enough the source), and local caching.

Hotlinking from the image server will reduce the load on the PQ as there will be no need to wrap the images into the GPX/zip file.

While I'm no network engineer, I'm thinking if there is a caching proxy in your route and it has the files needed it will serve them up instead of hitting the gc.com server where they are stored.

Lastly, a well behaved, off-line database like GSAK could cache the images locally so they would only have to be downloaded once. Even slicker would be an ability for GSAK to redirect the IMG tags inside the description to point to a local copy of the image.

Hotlinking to images and storing them locally is probably the best route. The best that I'm thought of, anyway.

October 25, 2005

But for those PQs that are turned into offline databases, I assume that the total bandwidth for retrieving the images is the same, irrespective of the the eventual timing

Not if the pocket query is run more than once, surely? Consider someone who has a daily PQ of the 500 nearest caches to them. One of the caches has a spoiler, plus 24 other photos, averaging 50kb each, taken from a camera which was left in the cache. In a year, including the photos in the cache in the PQs would take 25*50*365KB, which is around 450MB for that single cache. Hotlinking, on the other hand, would download about 50kb in total.

There are two main reasons why the T&Cs ban automata from accessing GC.com. The first is that, since GC.com have no copyright on submitted materials, they must protect themselves from data theft by T&Cs. The second is that they need to protect their servers from overload. Only if everyone downloaded every single cache photo every day, would the URL option put as much strain on the servers as the binary option. In the above, not particularly contrived example, it's about 9000 times more bandwidth efficient!

Edited October 25, 2005 by Teasel

2.7k · October 25, 2005

CoyoteRed Posted: Oct 25 2005, 01:59 AM
(5¢ @ Oct 24 2005, 09:54 PM)
(Waldo's Revenge @ Oct 24 2005, 07:34 PM)
(Black Dog Trackers @ Oct 24 2005, 05:47 PM)
What/where is the "current GPX specificaiton"?

Does it apply to benchmarks? If so, I might have some suggestions of things to add.

I agree can anything be done with gpx files for benchmarks?

I will third that one

I vote against it.

There is already a way to ceate GPX files for benchmarks. All you got to do is ask around.

Go HERE to get started.

Hope this helps.

No, it does not help. I assume you mean to get all the data from the NGS instead of GC.com. I'm interested in getting the data from GC.com to see what's been found, when it was found and especially what hasn't been found in GC.com logs.

7.2k · October 25, 2005

CoyoteRed Posted: Oct 25 2005, 01:59 AM
(5¢ @ Oct 24 2005, 09:54 PM)
(Waldo's Revenge @ Oct 24 2005, 07:34 PM)
(Black Dog Trackers @ Oct 24 2005, 05:47 PM)
What/where is the "current GPX specificaiton"?

Does it apply to benchmarks? If so, I might have some suggestions of things to add.

I agree can anything be done with gpx files for benchmarks?

I will third that one

I vote against it.

There is already a way to ceate GPX files for benchmarks. All you got to do is ask around.

Go HERE to get started.

Hope this helps.

No, it does not help. I assume you mean to get all the data from the NGS instead of GC.com. I'm interested in getting the data from GC.com to see what's been found, when it was found and especially what hasn't been found in GC.com logs.

Sure. As long as there is a seperate membership for that along with dedicated servers so it won't be bogging down the geocaching PQs.

Otherwise, no thanks.

October 25, 2005

Not if the pocket query is run more than once, surely? Consider someone who has a daily PQ of the 500 nearest caches to them. One of the caches has a spoiler, plus 24 other photos, averaging 50kb each, taken from a camera which was left in the cache. In a year, including the photos in the cache in the PQs would take 25*50*365KB, which is around 450MB for that single cache. Hotlinking, on the other hand, would download about 50kb in total.

But only, if the images are locally stored. If I view the page with the hotlinked images while I have Internet connection, then they are normally retrieved from the server - unless there is a local or proxy cache. And each time I hit the page with the image (the HTML may be local but the image is on the server), the images' kilobytes are downloaded from the server - at least that's my understanding of hotlinking. So, if I transform my local cache description GPX data with hotlinked images into an offline database, the images are (either taken out of a local cache or) retrieved from the server. This creates the hits and the bandwidth, IMHO. Hence, the bandwidth argument depends on if I do this less often than having PQ being sent to me.

The local cache of images acts quite the same way as a PQ would that is configured to include only caches that have been updated (does that include changes in the images, if the URL does not change?). If something changes on the server, it will be included in the transmitted data. Images of unchanged caches would not be included in the PQ as the cache would not be included. (This brings up the request for an attribute "that has not been updated since the last PQ that included it".)

Everything contained in a PQ including the images would IMHO be a clean way to get the images to the cacher. Hotlinking can create even more traffic, if I don't have a cache locally or at a proxy because everytime I hit the page the image is downloaded - and that can be more often than receiving a PQ.

But I agree, a fine way to do it would be to teach tools like GSAK to retrieve the images that are referenced in the GPX description of a cache once and to use the local copy happily ever after.

Best regards,

HoPri

October 25, 2005

Image caching brings up one other piece of data that I don't know exists or not.

The date the cache page was last changed.

This would easily let other apps know it they need to do anything new to the cache page or just use it from memory.

7.2k · October 25, 2005

Hotlinking can create even more traffic, if I don't have a cache locally or at a proxy because everytime I hit the page the image is downloaded - and that can be more often than receiving a PQ.

Only if your browser is not caching. Most do.

I think the argument is moot anyway. The URL/hotlink solution is a lot more viable than sending down images with the PQ. Most likely everything will be cached locally, folks will crunch their own thumbnails, and we'd still have the same end result all without the tremendous load on the PQ servers.

1.4k · October 25, 2005

CoyoteRed Posted: Oct 25 2005, 01:59 AM
(5¢ @ Oct 24 2005, 09:54 PM)
(Waldo's Revenge @ Oct 24 2005, 07:34 PM)
(Black Dog Trackers @ Oct 24 2005, 05:47 PM)
What/where is the "current GPX specificaiton"?

Does it apply to benchmarks? If so, I might have some suggestions of things to add.

I agree can anything be done with gpx files for benchmarks?

I will third that one

I vote against it.

There is already a way to ceate GPX files for benchmarks. All you got to do is ask around.

Go HERE to get started.

Hope this helps.

No, it does not help. I assume you mean to get all the data from the NGS instead of GC.com. I'm interested in getting the data from GC.com to see what's been found, when it was found and especially what hasn't been found in GC.com logs.

I think the current iteration of GPX seems to be OK for benchmark data downloaded directly from the NGS database. What I really want is the "Past Logs" feature.

I agree that a Pocket Query giving us a benchmark GPX would be desirable. Just add an additional checkmark for nearest BMs, and use the future <wp> field to obtain the user-posted adjusted coordinates. I'm willing to wait for TPTB to figure it all out, including the potential server load issue.

For now, it's probably off topic. Someone (like me) will probably start a new thread on this when appropriate.

2.7k · October 25, 2005

I never thought of it before, but apparently there's some competition between geocaching and benchmarking such that those of us primarily benchmark hunters should be barred from having the same benefits of the Site as geocachers based on geocachers voting no.

I wonder how much 'bogging down' the benchmark side would cause just to have a slightly larger number of fields when you check the boxes on a proximity list for a download. I'm not even talking about a pocket query exactly, just to be able to download the found status information on a list of benchmarks.

7.2k · October 25, 2005

...apparently there's some competition between geocaching and benchmarking...

Not in the least. It's just that this site is primarily for geocaching and the resources are thin as it is. I'd say the vast majority of the PMs here are here for the geocaching and would be terribly upset that functions for benchmarks slow the site to a crawl. I know I would.

2.7k · October 25, 2005

Well I can see that, but if x% of PMs are primarily benchmarkers and they cause (with as good as services as the geocaching side gets) x% of the server load, then what would be unfair about that?

What does seem unfair (in my complete quantitative ignorance) is that x% of the PMs should not be able to download find information At All.

Edited October 25, 2005 by Black Dog Trackers

7.2k · October 25, 2005

Eh, I don't see a problem as long as your total downloads aren't any more than anyone else's. :laughing:

Of course, I'm not sure how helpful that would be considering you'd probably burn through your 500 in just a few miles.

9.2k · October 25, 2005

Only if your browser is not caching. Most do.

I think the argument is moot anyway. The URL/hotlink solution is a lot more viable than sending down images with the PQ. Most likely everything will be cached locally, folks will crunch their own thumbnails, and we'd still have the same end result all without the tremendous load on the PQ servers.

You're right. Send the URLs of the images, but not the images themselves. This is already being done with in-line images. When a GPX utility, like Watcher, uses the GPX data to show the cache page info, it retrieves the image from the server.

I don't think most people have an idea of just how much this would bloat the PQs. To be embedded in a GPX file, an image would have to be converted into BASE64, or some similar ascii encoding scheme. As an example, I took the 348×274 image from this cache page, and converted it to BASE64. It's a jpg file, so it was already compressed. Here's the code that would be included in a PQ, if the file was included. Remember, this is just ONE not-very-big image. I apologize ahead of time for the size of this post, but I think some people just won't get it any other way.
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Edited October 25, 2005 by Prime Suspect

October 25, 2005

you know what they say, "A picture is worth a thousand words"

October 25, 2005

What does seem unfair (in my complete quantitative ignorance) is that x% of the PMs should not be able to download find information At All.

Well, if you're in pursuit of quantitative fairness, consider that of the 217 countries for which GC.com lists geocaches, benchmarks are not suppported At All in 216 of them :laughing:

Edited October 25, 2005 by Teasel

October 26, 2005

Some final remarks from my side, as we all do agree on the final goal and it's not our decision how exactly to achieve it anyway...

Hot-Linking is more efficient as long as local caching is done. Most people I know have their browser caches configured to be cleaned up after a session. Most personal firewalls and privacy service software cleans it upon logout. Proxy caches regularly kick out files that are hit only with a low frequency (as compared to other www files). Hence, local geocache image storage must be specifically accomodated for by special tools such as GSAK. The bandwidth efficiency is NOT a given of the hotlink method per se. If I hit refresh several times a day on a page containing hotlinked images, they'll be retrieved FROM THE SERVER just as often. The cummulative bandwidth consumed for just one image by this can be much higher as in a PQ, as PQs can come just once a day at the most. And, last but not least, we must not forget about the TOU issue. TPTB would have to grant the right to automatically retrieve images for any geocache database size I choose to built up by PQs. If automatic retrieval of hundreds or thousands of images at click-time is not "legalized" then it's useless. Local caching does not avoid this problem as I haven't even looked at most of the geocache descriptions on my PC when I transform them into an offline database. Hence, at that time and at least once initially, the images need to be retrieved from the servers. Afterwards, I agree, a local caching system would reduce the required bandwidth - just as it would be reduced if I configure my PQs to just include recently updated geocaches.

Nobody said the images would have to be embedded in ascii form into the GPX file the way that it was ingeniously illustrated above! Put the jpgs into a separate zip-file and include it in the PQ's zip. The GPX file size does not change that much at all, but the mail size in total does, of course. Clearly, zipping would not squeeze much more juice out the image files as they are already heavily compressed being the jpgs they are. At least to me, the zip method would have the additional advantage that no tool needs to be adapted. Groundspeak includes the images in a zip, the PQ mail files are bigger but I don't need to change software at home (ok, cachemate would have to add the image feature, anyway, but GSAK, for example, would not have to change at all). There is no TOU issue, and so forth.

I think the TPTB will have some more issues to think about here, such as how to best control the resulting traffic, how to ensure that it can be restricted easily to PMs (assuming that is what they want) and many more. Hence, we can only look forward to their decision as we all agree that we want the images somehow - one way or the other. EDIT: And rather sooner than later...

Best regards,

HoPri

Edited October 26, 2005 by HoPri

7.2k · October 26, 2005

Well, you've made your case and I've made my case.

I do think you've made a few wrong assumptions on browser caching, the load on the PQ server vs. the image server, concentration of the server loads, size of the files (huge your way,) and any TOU issue (there isn't any.)

Let the chips fall where they may.

October 26, 2005

Well, you've made your case and I've made my case.

I do think you've made a few wrong assumptions on browser caching, the load on the PQ server vs. the image server, concentration of the server loads, size of the files (huge your way,) and any TOU issue (there isn't any.)

Let the chips fall where they may.

Being the ever-learning person I am, I'd love to be educated by you about where I'm getting something wrong (in particular browser caching). It is always good to learn to understand things better. I suggest we do this by PM rather than in public here. I'm looking forward to your PM (if you will)!

Best regards,

HoPri

7.2k · October 26, 2005

I'm looking forward to your PM (if you will)!

It's in your box!

9.2k · October 26, 2005

Most people I know have their browser caches configured to be cleaned up after a session.

I would say that you know a very select group of people who do not represent the majority.

The default for most browsers is to keep cached files across sessions, because it speeds up the browsing experience. Files are not removed until a preset storage limit is reached, and then they are usually selectively removed based on date of last access.

Most people do not use the 'paranoid" settings, unless they have some real need to do so.

1.3k · October 26, 2005

Of course, I'm not sure how helpful that would be considering you'd probably burn through your 500 in just a few miles.

Well, since you could actually get up to 2500 a day, and there's a reduced need for updating the info at short intervals, I can see where this would be VERY useful to those of us who hunt benchmarks.

Also, it would be nice in the future to be able to get caches, benchmarks and waymarks for an area combined in one PQ. For me, this would be especially useful on trips.

New Iteration Of Gpx

Recommended Posts