Jump to content

Wow! Super fast pocket query deliveries.


Olar

Recommended Posts

I have 4 regular queries that run every Friday. I usually get my two low numbered files within minutes just after midnight PST and the other two high numbered approx. 17 to 20 hours later. Today they all arrived by 3:20 AM. Thats a tremendous improvement. Did someone supercharge the server? icon_smile.gif

 

Cheers, Olar

 

"You are only young once but you can stay immature forever"

Link to comment

I was going to comment on this as well. I was up late last night doing surfing when I remembered I did not set up my queries for the day. I looked at the clock and it was a few minutes past midnight on the east coast. I figured I would put them in to run for Thursday (it was still Thursday on the west coast). If they ran fine if not I would just do them Friday. When I checked my email today they were there and it looks like they ran just as I was clicking the button to save them. Getting ready to set up the ones for Friday so see what happens there. I don't usually get mine until about 6 in the evening if I set them up the day before.

Link to comment

They now generate in the order that they were last generated, so...

 

If you're somebody that runs your pocket queries every day of the week, you have to wait until someone who only occasionally asks for a pocket query to get one before yours generates.

 

New queries for the day generate first, then the ones that haven't been run for a while, and so on...

 

One tip would be to select your queries to generate, say, Monday, Wednesday and Friday, instead of every day. That will get your query faster.

 

We'll eventually be splitting queries to two servers so they run even faster. I'm currently watching the queries and there are around 470 left to go for the day. Friday is the busiest day for pocket query generation.

 

frog.gif Jeremy Irish

Groundspeak - The Language of Location™

Link to comment

quote:
Originally posted by Jeremy (Admin):

If you're somebody that runs your pocket queries every day of the week, you have to wait until someone who only occasionally asks for a pocket query to get one before yours generates.


 

This sounds like a good idea. I do most of my geocaching on the weekends and have my pocket queries delivered on Thursdays. They are always there on Thursday morning, giving me enough time to plan a dozen or so caches to hunt.

 

I only request 100 caches out of the possible 500, this gives me plenty of choices for the weekend hunt.

 

I'm glad to know my access method gets me the caches when I need them, and also glad that I won't have to wait in line behind those who download the maximum on a daily basis. After all, what's the point of downloading 500 caches each day, unless you are hunting caches on a daily basis?

Link to comment

quote:
Originally posted by cachew nut:

After all, what's the point of downloading 500 caches each day, unless you are hunting caches on a daily basis?


Because some of us keep our own database of caches. I like to keep a current list over a wide area because I never know where I might be heading or when I might be doing it. I also have set up an entire system of databases that work the way I want to work them. Kinda like ClayJar's program only not as nice and without a lot of the features he has. But I can work that data and drop it into other programs that I use very easily. It works well for me but maybe not for everybody else. So I grab the maximum amount of PQs and results I can everyday to keep my lists updated.
Link to comment

quote:
Originally posted by Jeremy (Admin):

...New queries for the day generate first, then the ones that haven't been run for a while, and so on...

frog.gif Jeremy Irish

Groundspeak - The Language of Location


So it appears under this procedure to run a one time query it would be better to create a new one and delete it when you get the results instead of keeping a query with no run dates then activate it when needed.
Link to comment

quote:
Originally posted by GrizzlyJohn:

Because some of us keep our own database of caches. I like to keep a current list over a wide area because I never know where I might be heading or when I might be doing it. I also have set up an entire system of databases that work the way I want to work them.


 

Ok. It just seems like a lot of work collecting and maintaining all that data on the chance that you might be in the area of a cache. Why not just pick a few from the list and plan on finding those?

Link to comment

quote:
Originally posted by cachew nut:

Ok. It just seems like a lot of work collecting and maintaining all that data on the chance that you might be in the area of a cache. Why not just pick a few from the list and plan on finding those?


It is a bit of work. But I am a geek so it has also allowed me to play with some programming ideas. Sometimes I just never know. I did that once and then ended up someplace different with no list of caches to go after. I vowed never again. But really at this point I have a ton of it automated so I don't spend that much time on it and I sleep better at night knowing I have a complete list. :-)
Link to comment

quote:
Originally posted by Jeremy (Admin):

They now generate in the order that they were last generated, so...>snip<

New queries for the day generate first, then the ones that haven't been run for a while, and so on...>snip<


 

That's a great algorithm. Gives 'impromptu' queries a quick response and encourages folks to run regular queries judiciously.

 

I like it! Thanks for the fix!

 

If the automobile had followed the same development cycle as the computer, a Rolls-Royce would today cost $100, get a million miles per gallon, and explode once a year, killing everyone inside. -Robert X. Cringely, InfoWorld magazine

Link to comment

quote:
Originally posted by HartClimbs:

That's a _great_ algorithm. Gives 'impromptu' queries a quick response and encourages folks to run regular queries judiciously.


I think it kind of says that 'impromptu' queries are more important that regularly run ones. I don't think that seems right. I think they are all of the same importance. When I became a member I was promised 5 queries a day. Not that I would have to run them judiciously. That is a part of the membership I paid for, an offer was made and I accepted it. As long as it does not effect me getting my queries that it seems OK but I still don't like the idea that some queries seem to be more important than others. I was asked earlier here why some people would want to run a query for 500 caches. I think it might also be fair to ask why some people can't plan ahead. Not flaming anyone its just that we all do things a little different.

 

Does anybody have any idea how many queries on average are run everyday?

Link to comment

I agree that it isn't a good idea for some queries to be more important than others. However since they can't be run simultaneous, some are going to run before others. The old method run the oldest queries i.e. the ones with the lowest numbers first. Now per Jeremy they run "order that they were last generated". Is either method completely "fair", probably not. Is there a completely fair method, probably not.

Link to comment

quote:
Originally posted by AllenLacy:

Is either method completely "fair", probably not. Is there a completely fair method, probably not.


I think first one in first one out is fair. We all have to wait in line from time to time for something. I never complain about the person way up ahead of me in line because they got there before me. This new system seems to encourage butting up in line. Like I said I don't mind waiting for my queries to arrive. But there have been many days in the past I did not get my queries because from what I have seen they just run out of time and do not run. But if I did them on the spur of the moment it seems like they would run under this plan. But there is the suggestion of not editing ones queries, to instead just delete them and create new ones. I will probably experiment with that later today or tomorrow.

 

There are many things that can be done to make queries run better. Many have been suggested here in the forums. And I know I have suggested some off line to TPTB. But I have not seen any of them take place. As an IT person when I hear more hardware being thrown at the problem that sends up red flags to me.

Link to comment

Some people are more equal than others...

 

It was a tradeoff. One, I wanted to give people who have been here since the beginning the benefit of first queries each day. So I started to order queries by their ID numbers. First in, first out.

 

However, people who wanted to go for a quick jaunt on Friday morning need to get their queries faster because they are "in the mood." So we need to accomodate them as well.

 

Considering that pocket queries now all run almost before noon most days (except Fridays where there are double queries), it seemed to be the best method. Both still get their queries but folks that are "in the mood" will get them faster.

 

It is true that you could remove your queries and create new ones every day, if you wanted. But that's fine since it would be assumed that you are one avid geocacher to do this.

 

It's a behavioral decision, and it seems to be the "elevator music" solution where no one likes it but it isn't totally unpleasant.

 

frog.gif Jeremy Irish

Groundspeak - The Language of Location™

Link to comment

I'm still curious about running the pocket queries on a daily basis. I guess I'm not understanding something, which wouldn't be the first time.

 

I'm not trying to put anyone down, I'm just trying to find out that missing piece of information that is eluding me.

 

Let's assume 2 examples, then maybe someone can point out the difference.

 

Example 1:

You get the maximum pocket query, each day, seven days a week, but do not go geocaching all week, just use this data for a personal database.

 

Example 2:

You get the maximum pocket query, perhaps every other day, or maybe 3 times a week, but do not go geocaching all week, just use this data for a personal database.

 

Questions:

1- On day 7, how would the database in example 1 differ from the one in example 2?

 

2- Would the pocket queries come quicker if more people followed example 2 over example 1?

 

3- Isn't the primary function of the pocket query to assist a geocacher who wants to find caches?

Link to comment

quote:
Originally posted by cachew nut:

I'm not trying to put anyone down, I'm just trying to find out that missing piece of information that is eluding me.


In your examples are you thinking that the same queries are being run everyday? If so that may be the missing piece of information. The queries I run everyday are different. I don't want to go into the entire explanation of how I have things set up here. But if you want to know just let me know and I really would be more than happy to explain all of the queries that I run and why. But I will say that if I could query for archived caches I would only need a couple of queries a week and not five a day everyday. My guess is that most people would not do things the way I do and that is OK. But I really don't want to post a long description of how I do things for people that would not be interested or care. Even a simple post from me has a way of getting wordy. So if you are still scratching your head please let me know, I can explain it all to you. In the end, you may still wonder why someone would do things that way but I think you will understand why someone could get to the point of doing them that way.
Link to comment

quote:
Originally posted by Jeremy (Admin):

Considering that pocket queries now all run almost before noon most days


Ok this is all a fair run down of how you came to this. And it does all work for me as long as at some point I will receive the queries I have scheduled for that day. If I get them then I don't really see the need to delete and create new queries instead of just editing the ones I have there and keep the same ID number.

 

The only question I have is for the most part I don't get my queries until after 5 or 6 in the afternoon sometimes later on the east coast, 2 or 3 o'clock your time. I really have not looked at the headers on the email but is there a delay between when the queries are processed and when they actually get emailed? If not then there may be a delay somewhere. It may be on my end or maybe on yours or somewhere in between. But if you are expecting most queries are emailed by noon your time and they are getting delayed on your end by a long period of time would you want to know that? Again I really have not looked at the headers to see how they move but really once it hits the email system for the most part it is in the person's mailbox almost at the same time. Yes the attachments can be large so that may slow it down some. But I will try to remember to take a peek at the headers and see if they are getting delayed somewhere.

Link to comment

quote:
Originally posted by GrizzlyJohn:

In your examples are you thinking that the same queries are being run everyday? If so that may be the missing piece of information. The queries I run everyday are different.


 

Yes, I assumed that they were all the same. But if they are not the same, then your data is only as current as the last data collected. If all you are trying to do is find out which caches have been archived, rather than having the most recent that pop up, then I agree that it would be nice to have a list of archived caches that are accessible. There has got to be an easier way to keep track of archived caches.

Link to comment

quote:
Originally posted by GrizzlyJohn:

I really have not looked at the headers on the email but is there a delay between when the queries are processed and when they actually get emailed? If not then there may be a delay somewhere. It may be on my end or maybe on yours or somewhere in between. But if you are expecting most queries are emailed by noon your time and they are getting delayed on your end by a long period of time would you want to know that?


Pocket Queries are sent immediately once they're built. For example, the average delivery delay for today's queries was about 8 seconds. So you should be getting them right away. If you think there's a significant delay somewhere, let me know and I'll look into your particular queries more closely.

 

frog.gif Elias

Link to comment

quote:
Originally posted by Elias:

Pocket Queries are sent immediately once they're built. For example, the average delivery delay for today's queries was about 8 seconds. So you should be getting them right away. If you think there's a significant delay somewhere, let me know and I'll look into your particular queries more closely.


Thanks that is good to know. My queries today arrived just a few minutes after 3 in the afternoon my time. They hit my mailbox a minute or two after coming out of your mailbox. My only point on that is if indeed you were expecting that most queries were done by noon that was not what I was seeing. Of course the first day I was going to check they arrived right on time. But unless somebody says something you may never know there is a problem. But if I see them starting to arrive late regularly I will try to check and see where the delay was coming from.

 

Thanks.

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