The Erfworld Forum

[Elm11]

Hello. I'm just a little inquisitive, but will the Erfworld forum be closed down now that erfworld is moving to foreign shores? Just wondering, and i don't think this warrents it's own thread, but i couldn't find anywhere else appropriate to put it. :smallsmile:

[NerfTW]

From the Erfworld sub forum-


For the moment, nothing.

Unfortunately, there's no way to transfer posts to the new Erfworld.com forums, so eventually, this board will just be locked. But not for a few months at least. After that, it will probably exist in a read-only fashion until we take down the Erfworld comics, sometime in 2010 or so.

[Zherog]

There's actually conversion tools out there. I found this one with a quick google search (http://www.phpbb.com/community/viewtopic.php?t=54908), although I'm quite certain it's not the correct versions of vB or phpBB.

The worst part of any conversion would be usernames. I'm sure there's people who posted in the Erfworld forum here who haven't registered over there. If there were some clever way around that, though (such as putting them all under one user and putting the original use name at the top of the post) it probably wouldn't be terribly difficult to write the conversion. Both databases use mySQL, and I believe both have similar schemas.

edit to clarify: I've re-read my post, and I might be giving the wrong impression. I don't think a conversion would be easy; but I do think it'd be possible.

[Rawhide]

Zherog: Your post gives the impression that it is possible to extract the threads and posts from a forum and insert them into another forum's database. I would just like to confirm that no, it is NOT possible in any feasible way. This would not be at all like a simple* database conversion.

*Database conversions are not simple at the best of times and are fraught with difficulties.

[Zherog]

It certainly is possible to do. Writing the query to pull the data out of vB is pretty simple. The hard part would be getting the data back into phpBB. In my opinion, even that wouldn't be an impossible task.

Now, that said... I indicated in my post that it wasn't easy. Maybe you replied before I added my edit. The real question isn't whether it's possible (because it is). The real question is whether the effort involved in doing so would give enough of a benefit to make it worthwhile. The answer to that question is almost certainly, "No."

[Renegade Paladin]

I've seen conversions between boards done. These were, however, entire boards, not a particular subforum. I also have no idea what actually goes into it; just that it was done.

It seems to be that simply transferring the posts without taking the entire GitP database with it would have major problems with post attribution even if everything else was easy.

[Zherog]

Yep, most definitely. Doing the whole thing would certainly be easier -- again, in large part because tools for that are already out there and available. Taking what amounts to a small subset of the data most definitely complicates things.

Is it possible to do? Yep. Is it worth the effort involved? In my opinion, no.

[Rawhide]

Is it possible? Anything is possible if you want to argue philosophically about the meaning of possible. You will note that I specifically said that "it is NOT possible in any feasible way".

It's about as possible as NASA landing someone on Pluto this year.


P.S. Database conversions taking the entire database from one format to another are possible, but far from easy or simple.

[Zherog]

It's about as possible as NASA landing someone on Pluto this year.

You're way over-exaggerating the complexity of the issue.

1) Write a SQL query that takes all the post data for the specified forum.
2) Execute the query against the vB database and send the data to a flat file. Comma delimited is preferred, but not necessary
3) In the mySQL database, create a custom table that has one column for every field in your extract file.
4) Load the flat file into your custom table.
5) Write another SQL script to take data from your custom table and insert records into the phpBB tables.
-----A) Given that users don't exist, the best solution for that would seem to be to create one user named OotS Forums and assign all these posts to that user.
-----B) Using text manipulation strings in SQL, the original poster's name can be added to the post, either at the very beginning or at the very end


This sort of thing is far from rocket science. I write one of these about once a year. This one has the benefit of having a relational database on both ends, unlike some of the ones I've had to write. (Lotus Notes databases suck. And I use the term "database" very loosely there...)

It's quite possible. I have no doubt that I could write it. I don't think I have any super-special SQL programming skills, which leads me to believe another developer could also write it.

The problem isn't whether or not it could be done. The problem is that the "cost" (even if I did it for free, it still costs me time; same is true for another developer) isn't worth the benefit derived from the task.

[Rawhide]

And yet, you've completely overlooked the hurdles and artefacts that will be encountered and left in the extraction and reinsertion process. It is not as simple as you are making it out to be and the results won't be as peachy as you claim.

I can name one huge obvious flaw in your plan right now without going into much detail. Who wants to read through thousands of posts that don't identify who the poster is/was? Who wants to read through a conversation involving 5+ people who have posted more than once each without being able to determine who is who?

It's not, in any stretch of the imagination, a simple task. The data would have to be transfered securely without affecting the privacy or security of our members and member database and the data has to be useable, legible and readable in a meaningful format on the other end.

Many strange errors, glitches and bugs that have crept into a database this size running on hardware that was not adequate will have to be taken into account/consideration in any extraction and reinsertion. Error handling code would need to be developed and a manual inspection would be required of the transferring data.

I can go on...

[Zherog]

>>snip<< for space

I've already identified username as one big flaw that would need to be considered. In fact, I considered it way back in my first post in this thread. Short of importing the entire user table (which is pointless, even before considering privacy issues which also makes it a no-go), that would absolutely have to be something that would get worked out. My real quick outline provided just one potential solution to that problem -- and I didn't spend more than 10 seconds thinking about it.


I can go on...

Please do. :) Everything you've listed is stuff I've already considered, and part and parcel for writing this sort of transfer.

And don't misunderstand me. I'm not saying it's a cake walk of an assignment. It's not. But it's certainly not on the same level as "NASA landing someone on Pluto this year."

I do this sort of task at least once a year in my job. And it's always with data that's more important than post data, such as open Accounts Receivable records, General Ledger balances, unpaid Accounts Payable transactions, Accounts Payable payment histories, and so on. I guarantee you those are harder, if for no other reason than the fact of dealing with auditors who insist on balancing to the penny. Crazy auditors. :)

And I'll say it yet again: just because it's possible to write it doesn't mean it should be done.

[Rawhide]

Ah, I see what's gone wrong here. Regarding that one statement about NASA, you're arguing that the technical knowledge is less complex (i.e. no one will die if we get it wrong), whereas, I'm arguing that the possibility of it ever happening is about the same (i.e. NASA has thousands of technical staff that will work on any space launch with lots of time, whereas, we don't).

[Kroy]

Ummm..., how much does it really matter? I mean, one of you is saying that it would be hard and not worth it, while the other is saying that it is almost technologically impossible with the resources of Giantitp. Can't we agree that it's not going to happen?

[LordZarth]

A good friendly debate/argument/discussion is stimulating.

...but seriously, by asking that question you show that you obviously aren't one of those people. FYI, there are lots of people (including me) who will debate/discuss for no particular reason, just like chatting.

[Renegade Paladin]

Ummm..., how much does it really matter? I mean, one of you is saying that it would be hard and not worth it, while the other is saying that it is almost technologically impossible with the resources of Giantitp. Can't we agree that it's not going to happen?
Where's the fun in that? :smallbiggrin:

[Zherog]

OK, so when you said...


It's about as possible as NASA landing someone on Pluto this year.

I indeed thought you were taking about the theoretical possibility of somebody being able to write the conversion. And that's why I disagreed with you, because I knew it was possible to write a conversion.

But, based on your most recent post that's not what you meant. You weren't talking about the theoretical possibility; you were talking about it actually getting done.

Well, yeah. I already sad - several times - that the cost-to-benefit is too low to make it worthwhile for anybody to write it.

[Elm11]

thanks for answering my question guys. :smallsmile:

[Zherog]

You probably got more than you bargained for, eh? :smallwink:

[Skorj]

As I often tell my boss "in software, anything is possible, but you probably care about the cost".

Writing a Python program to simply read all the posts in one forum and post each to another forum using appropriate, newly-created usernames (no database access at all, just get and post as normal users) would take a talented guy a day or three, if the source forum were locked down. But I'm not sure why anyone would bother.

[Elm11]

You probably got more than you bargained for, eh? :smallwink:

Yes, yes i did :smallamused: