View RSS Feed

WoW-GPS Dev Blog

WoW-GPS 2.0 - Filling the TUJ Void?

Rate this Entry
To be (a TUJ replacement), or not to be (a TUJ replacement) - is that the question?

I'm sure by now, everyone is well aware of the technical issues the The Undermine Journal is facing and the site is at risk of shutting down for good. If you're not, then please let me know in the comments why you're reading this blog, as I'm very confused as to how you got here

Due to the magnitude of the Undermine Journal project over these past few years, any extended absence will obviously leave a considerable void in the gold-making community. As the primary source for item data/history/pricing, seller data and also for AH notifications, TUJ has easily become most goblins' #1 resource.

Based on how the WoWGPS project is moving, we're technically not too far away from having access to a full 2 weeks of API data, which is more or less what TUJ works with, so I think it naturally begs the question: Could WoW-GPS serve as an effective replacement for TUJ, both with data (item and seller) and with notifications?

I hadn't ever had any plans to offer pricing notifications through WoW-GPS, at least not until I had something unique to offer (ie. complex notifications, such as shuffle profitability thresholds, etc.) Part of the reason has been because the functionality already exists. This point is obviously in a bit of limbo for now, but there are still other alternatives, and I'm also aware of the efforts of @Sapu94 and the TSM team to restore the deal notifications to the desktop app, so I feel like it would be a wasted effort to replicate existing works. There's also the technical hurdles - The data server is running smoothly right now, but I'd be more than a bit nervous to add additional bulk to the processing scripts and risk breaking core functionality. There are obviously a few different ways to implement notifications, but each comes with it's own setbacks. If I piggyback on existing processes, I risk draining resources and unnecessarily bloating them. Another option would be to create something that interfaces with the Data Server AFTER the major processing has been done, but the drawback here of course would be the additional delay in timing. I think the cleanest solution would be something that doesn't actually store the API data at all. I could theoretically run 2 heroku apps (since I can only have 1 worker OR 1 web process in order to remain on their free tier) which would both manage the notification settings (users, realms, notifications, etc.) and scan through the API files. The drawback here would be twofold: 1) The app would essentially exist in isolation, so there wouldn't be any quick way to add notifications from item or seller pages the way TUJ does right now. 2) Heroku's free DB only has capacity for about 10,000 rows (their estimate), so I probably wouldn't be able to offer a free, open-ended service through this method. I know that TUJ is currently running as a paid service for notifications, but it's also been operational and tested for quite some time, I'm not sure I'd feel right about throwing a subscription fee on something that was hacked together in a weekend (or so...). I'd feel obligated to actively work/develop/support the project until it was stable, which would inevitably pull me away from the rest of the WoW-GPS project - not something I'm too keen on. So, notifications are likely not going to happen, especially if they can't be tied to anything other than raw API data.

As for the item/seller pages, the short version is that it was probably going to end up in the new WoW-GPS model at some point, anyway, but it also was probably going to look a lot different than TUJ as well. WoW-GPS is positioned to be quite a bit more than simply price data/history. WoW-GPS will also have more detailed information about item sources, such as crafts, etc. and most importantly, the shuffle yields. It will also have user-specific item information which will likely result in a different sort of UX taking place that what makes sense for TUJ and then there's the fact that more detailed seller/competitor analysis is something that's prety high up on my own feature list anyway - I don't think the WoW-GPS data pages would really mirror TUJ's, so I'm not sure that it makes sense to rush them into production, posing as a replacement. WoW-GPS will almost definitely have item stat pages at some point, but I'm not sure they make sense this early on in the development cycle, so I'm more inclined to focus on other aspects, first.

I'm also constantly aware of the technical scope of the project, and while I believe my VPS (where the public-facing application will end up residing) can handle things without issue, I'm cautious of putting too much strain on the data server until I know it can handle it. I'd also want to prioritize the core WoW-GPS features (shufflers, professions, etc.) and ensure that the Data Server is capable of serving their functionality before looking to expand. Based on the limits that I've already reached with the Data Server (see previous few blog posts and ensuing comments), I'm not 100% convinced this will serve as a long-term solution just yet.

So, TL;DR - WoW-GPS could potentially offer some TUJ-replacement functionality, but almost definitely not before the shutdown at the end of the month, and likely not anytime soon, anyway. On top of all that, I'm not yet convinced it's worth the rush/extra effort, so I'm more inclined to hold off and gradually implement TUJ-like features into the WoW-GPS project where they make sense, and as a compliment to what I believe is more fundamental to the aims and goals of the project.

Convince me otherwise and we'll talk


  1. Kathroman's Avatar
    I changed my mind while thinking about it on my walk to work: