EVE Technology Lab

Like any user supplied data even the API data could be faked from the client but like you said it would make more information available which is a good thing and can be used to help detect the bad data. One reason I made the suggestion is I know at least in the case on Yapeal it spends a lot of time doing nothing but waiting on the API servers or the database server but it doesn't use much of the outbound network bandwidth so if people were interested I could look at adding it as an option. I can also see it being something any other applications that work with the API data could add. If all local applications that have marketing functions could easily add API and/or cache anonymous data uploading it would make a lot more data available and then the biggest problem is for the sites to handle the load

But... you know, when I start it it does exactly nothing. I wish I could get it to work. I do support your idea at the very least, especially since java is by definition multiplatform, which is awesome for an uploader!

BTW, sorry for not doing anything this weekend guys, RL is getting in the way.

Let's just say that EVE Marketeer is a production system, but that the production system is overloaded ATM. Fixes are under way, but they require a major rewrite of the codebase (meaning I'm busily writing now). In the meantime, EVE Marketeer is sometimes a bit erratic in terms of access, though it's never truly down. In the sense that it will eventually be accessible again without me having to reboot the server.

EVE Marketdata somehow always has a period of intermittent 503's somewhere around downtime. I think it has something to do with the database being regenerated.

I imagine there are some differences in the order of fields, but the concept is the same. If you have any requests, I'll be happy to implement them.

I've only ever done server-side programming, so I'm out of my depth with the uploaders, but having a universal uploader would be fantastic. I would immediately switch from Contribtastic to a universal uploader to help get data to the newer sites.

Ideally it would have a standard API for upload (or possibly change the order round). and be able to have new sites added via a config file. I know that I'd think about setting up my own personal site for the items and regions I'm interested it.

On a semi-related note, folks need to be spreading the word in their corporations / contacts that more people need to run the upload scanners (preferably one that feeds multiple websites). There's a lot of holes in the data on the various sites - even for regions in hi-sec. The Amarr trade-hub data is usually a few days out of date, places like Everyshore or Tash-Murkon are generally a week or more out of date.

(Personally, I try to make sure that Sinq/Heim are up to date at least every 2 days. But more redundancy would help a lot.)

So after looking at the format used by the above two sites I came up with an idea for how to layout the JSON version of the same thing. Here's a couple of example for everyone to look at:



The above have been formated for easier reading. To better understand the columns in "log" you'll probably need to have a look at page from EMK when it's back up or get it from repos and just open in as local file in browser. You can find it in tpl/api/display.html. I gave the examples as GET request but POST is much better. I made this just to get some comments and with a minimum of changes to see if anyone else had ideas how it might be improved on.

Your 'history' JSON has 'avg_price' listed twice in the header. I assume that's a typo, and the second one is volume.

Given I know no JSON whatsoever, it was still readable (much more so than XML, sorry guys).

Otherwise I support this effort, though I currently do not run an uploader (thought HQ2 had it already, but not checked all the new things yet).

@ Callean - My own guess for eve-central's performance would be that hardware was thrown at it. At my work the higher-end DB systems use RAM disks (not SSD, actual RAM-based external disks - connected by 2-4 FC links in Solaris Multipathing). Vendors like FusionIO are starting to break that mold for the lower-end (just would be nice if they had SPARC drivers on Solaris).

Just an odd thought. Since the data originates on the CCP servers, and since CCP has introduced a new, much more granular api system, and since they are currently in the mood to listen to the player base...

Why not start a concerted lobby for them to open the pricing api to the market information providers. There, one source of data, eliminates cache scraping, file uploads and it is accurate. And most of all, it is the primary and complete set of data.

otto