A brand new EN World in development

Morrus said:

The core of the new site will be an upgrade from our venerable vBulletin 3.8 installation to the more recent vBulletin 4 forum software;

Click to expand...

Does this mean someone will Pathfinderize the old version of ENWorld?

1. I am very disappointed that you won't be importing the database over. So many good posts down the drain, not to mention the fact that I'll have to sign up as a member again, lose my post count and my join date.

2. Do we really need a new edition so soon? I don't think this the current edition is broken, Morrus is just trying to force us to re-buy all our old posts.

Just think of the edition wars this will cause on the boards, and the fracturing of the EN World community.

Cries of "I prefer how moderation was handled on the old boards" and "I prefer the good old days of vBulletin 3.672 where my options were to quote or post. None of this fancy "Multi-quote" business or "Quick Reply" hootananies."

You really didn't think this through at all did you Morrus.













What?

Olaf the Stout

Geez, I just re-registered a decade ago. I hope they migrate everything over this time.

Will the new site have a pie dispenser...?

Sounds awesome, Morrus. Just one suggestion. Don't make the switch anywhere near the release of dndnext, please. I would really like to have access to ENworld during those days

Jack99 said:

Sounds awesome, Morrus. Just one suggestion. Don't make the switch anywhere near the release of dndnext, please. I would really like to have access to ENworld during those days

Click to expand...

It won't be that far off! Once we actually start work (and I believe we'll be making a start tonight) we'll be looking at a timescale closer to weeks than to years. I don't know how long it'll take, but I can't imagine it'll take all that long. Most of the time will be tweaking and testing.

Morrus said:

It won't be that far off! Once we actually start work (and I believe we'll be making a start tonight) we'll be looking at a timescale closer to weeks than to years. I don't know how long it'll take, but I can't imagine it'll take all that long. Most of the time will be tweaking and testing.

Click to expand...

Cool. That would work

Morrus said:

It won't be that far off! Once we actually start work (and I believe we'll be making a start tonight) we'll be looking at a timescale closer to weeks than to years. I don't know how long it'll take, but I can't imagine it'll take all that long. Most of the time will be tweaking and testing.

Click to expand...

Given the NUMEROUS discussions here on the nature of development, any estimate on time to delivery, no matter how conservative, should be tripled before being announced, and the number that is announced should then be doubled, and that number should then have a few months added on to it just to be sure.

So... in other words... in about a year or so

Kzach said:

Given the NUMEROUS discussions here on the nature of development, any estimate on time to delivery, no matter how conservative, should be tripled before being announced, and the number that is announced should then be doubled, and that number should then have a few months added on to it just to be sure.

So... in other words... in about a year or so

Click to expand...

In today's world, there is no such thing as a finished software product. Iterations are the current thing, and I think they can work here as well (I won't dwell too much on the theory of agile development, just try to give some practical advice):

1. Set up a fixed time period for an iteration. From my prior experience, 1 week is too short, 4 weeks is too long. 2-3 weeks is ideal, but the important thing is that each iteration must last the exact same amount of time.

2. Compile a list of stuff that needs to be done. Divide it into tasks. Each task should be fairly easy to implement.

3. Set up a goal for each iteration and then choose a number of tasks from the list that lead to that goal. The goal should be some functionality that's visible to the future users of ENW.

4. Resist the temptation to do tasks that are not on your "to do" list.

5. Make daily progress reports (internal or external).

6. At the end of the iteration, you should have come close to fulfilling your goal. Try to create a demo of sorts, or, better yet, deploy a new (alpha, beta, RC) version of the site for people to see and try.

7. As people test the site, new issues will arise. Add them to your list of tasks to do.

8. At some point, decide which features will have to be included for the new site to be rolled out. This is the Release Candidate (RC) version. From that moment on, institute "feature lock." No new features can be added (resist the temptation) until you've ironed out as many bugs as possible.

9. Deploy the new ENW, while continuing to work on the remaining features.

10. Remember, it's perfectly fine to roll out software that doesn't do 100% of the stuff that you think it should do. Google does this all the time.

11. Rolling the software out will result in the most valuable thing there is - user feedback. Analyzing user feedback will help shape the future of the project - perhaps you'll realize that one of your initial ideas won't be well accepted, but you'll get a different idea that users will appreciate more. Discarding old ideas and embracing new ones is natural in software development.

Hope some of this helps

Due to the whole importing the database thing, we will have to make a decision very soon.

Either launch sooner with only basic functionality, and re-add gradually; or have an extended (talking 3 weeks or so) downtime and launch fully functional; or have no downtime but lose 3 weeks or so of data added to the existing site while the new one I'd being worked on. The problem logically extends from the fact that the initial upgrade resets the database to 'factory settings' and once work begins on functionality if the new site, it alters the database structure, and so a later re-import of data becomes non-viable.

We've been analysing the work needed, and we're pretty confident it's about 3-weeks' worth. Luckily a lot if it is automated and is a case of reinstalling updated products. The only two real crunch points are the dice roller and Gamers Seeking Gamers, which were both custom coded and won't work as-is on the new software base. So that'll require finding a good and cheap coder to spend some time on it.