The biggest issue with the new Character Builder:

[wedgeski]

I've been a professional software developer for over 32 years now. I regularly lecture about software development processes and software testing. There is no way the thing launched on Tuesday would have gotten a "go" at any sane go-live meeting. Pushing garbage like that live tells me that WotC has absolutely no quality process in place. Or ignoring it completely.

All it tells me is that there are exigencies which the software developers and project managers of DDI have absolutely no control over, whatsoever, and which override any priorities they hold dear as to the state of the software they want to release.

Once upon a time in a planning meeting, someone somewhere decided the CB2 would launch with Essentials. Confidence was expressed by the DDI project manager at the time that such a deadline could be met... so much confidence, in fact, that no contingency was put in place for Essentials and/or Dark Sun support in the existing CB. Or perhaps it was simply a cost issue, we'll never know.

Now what has happened is that, for whatever reason, CB2 isn't ready for prime time (and for god's sake, buggy as it is, it's not the train-wreck this board would have you believe), but the schedule had to win. DDI support for Essentials was absolutely dependent on CB2, so it had to be released. Perhaps there was statistical evidence that people were actually changing their spending habits because DDI was lagging (and I mean proper evidence, not the deluge of threats to quit posted on boards like this), or perhaps some manager somewhere had staked his job on it.

We can argue forever about what went wrong (planning? feature creep? poor development methodologies? poor testing? simple bad luck because 10% of developer time was lost due to the 'flu?) but we'll probably never know the truth. Landing at a point where you're sure that the software team is inept is understandable, but unfair.

 

[avin]

Q&A for software programs is 50% part of my job at this moment.

CB wasn't ready to launch.

 

[Herschel]

I've been a professional software developer for over 32 years now. I regularly lecture about software development processes and software testing. There is no way the thing launched on Tuesday would have gotten a "go" at any sane go-live meeting. Pushing garbage like that live tells me that WotC has absolutely no quality process in place. Or ignoring it completely.

But let me just sit back in my armchair and watch this train wreck...

Spoken like every other software developer I've ever talked to. Now you know what the end user feels like. Developers always talk about how their product is better and they can do a better job, yet in the end, they're all about the same. That's not a knock, it's human nature, and there's a lot of good developers out there but end users' opinions of software is usually quite different than the developers' in my experience working with stuff for Fortune 5 companies.

I also know some times there are issues not with teh soiftware itself, but servers, etc. in the chain. Yet people, even IT people, jump on the software first (ie: system crashing).

Also, it's not like developers/programmers are in complete control of a project. I'd venture a pretty good hypothesis that the deadline was already pushed back once on teh project and they weren't given a second "extension". There are deadlines they are forced to meet.

I talked to my son who tried it opening day and he said he had one "crash", and he just refreshed and went on with creating some characters. No other problems.

Are there bugs? Sure. Were there going to be? Of course. It it surprising a few weren't found prior to release? That's a matter of opinion, there's a couple I was surprised about, but they appear to be already worked on. That said, the program isn't an "epic failure" or any of that other hyperbolic garbage, it has some bugs that are being worked out. It's the same stuff I hear about Microsoft products every time a new one is released.

 

[Herschel]

Silly me, using it to prep pre-gens for at least two Gamedays, my own character and one other person's character for two home campaigns, and numerous other trial-characters just for the fun of it. Guess I shouldn't have been able to use the thing to actually, y'know, GET GAMING DONE because it wasn't "functional."

LoL, so it worked for Essentials and Dark Sun characters? It only worked for "older" stuff, and if that's all you wanted, then yeah, it was good enough. But there was a whole new setting AND a whole new set of builds/classes/style of character needing support.

The shifting goal posts are pretty funny. Those screaming they have no confidence in WotC delivering useful digital tools and decrying the new character builder is an "epic failure" are the same ones claiming the update to the current version would have been just peachy.

 

[Jhaelen]

There is no way the thing launched on Tuesday would have gotten a "go" at any sane go-live meeting. Pushing garbage like that live tells me that WotC has absolutely no quality process in place. Or ignoring it completely.

Well, being a software developer myself, I have to say: go-live meetings are rarely sane.

If a software project is late (particularly a large one), it is often decided to release it prematurely, simply to get in some cash to be able to actually finish it. Promised features are cut from the initial release and the short respite gained is used to work on the first in a long succession of service packs.

Anyone remember Windows Vista? It was a similar situation. Windows 7 resembles what they'd _really_ wanted to release if they'd had been allowed more time.

 

[Danzauker]

Anyone remember Windows Vista? It was a similar situation. Windows 7 resembles what they'd _really_ wanted to release if they'd had been allowed more time.

Windows Vista is a perfect comparison to the state of CB as presented now.

Fortunately, we could keep Windows XP and go on with our business in my company, and we're only recently migrating to Seven...

Microsoft has the de facto monopoly of OS software on retail computers. And even they could not pass unharmed through the Vista fiasco. There's no way WotC can make me pay for crappy software. When and if CB will be ready, I could pay for it. I think it's reasonable and easy to understand.

 

[nerfherder]

Spoken like every other software developer I've ever talked to. Now you know what the end user feels like. Developers always talk about how their product is better and they can do a better job, yet in the end, they're all about the same.

My experience of working in the industry is that there is a huge span of skills and competence between developers, testers, managers, and all the other people that are involved in any IT project. I've seen projects succeed because the team was highly motivated, professional and focussed, and I've seen projects flounder because a manager wouldn't stand up to customers that argued amongst themselves and wouldn't commit enough of their time to agreeing requirements.

I also know some times there are issues not with teh soiftware itself, but servers, etc. in the chain. Yet people, even IT people, jump on the software first (ie: system crashing).

Whereas when I was talking to another IT professional just before the launch, we were debating whether it would be load on the database server or the network infrastructure that caused the biggest problems. We assumed that the software itself would be relatively bug-free, because that's the easiest component to test.

Also, it's not like developers/programmers are in complete control of a project. I'd venture a pretty good hypothesis that the deadline was already pushed back once on teh project and they weren't given a second "extension". There are deadlines they are forced to meet.

The most difficult part of making an IT project successful isn't the technology; it's people. Specifically, managing the stakeholders - from the WotC accountants to the customers. Your hypothesis is probably correct - and it's also probably more complicated than that.

 

[Festivus]

My experience of working in the industry is that there is a huge span of skills and competence between developers, testers, managers, and all the other people that are involved in any IT project. I've seen projects succeed because the team was highly motivated, professional and focussed, and I've seen projects flounder because a manager wouldn't stand up to customers that argued amongst themselves and wouldn't commit enough of their time to agreeing requirements.


Whereas when I was talking to another IT professional just before the launch, we were debating whether it would be load on the database server or the network infrastructure that caused the biggest problems. We assumed that the software itself would be relatively bug-free.


The most difficult part of making an IT project successful isn't the technology; it's people. Specifically, managing the stakeholders - from the WotC accountants to the customers.

I see this every day at my work. We have a team of developers who know Visual Studio very well, but ask them to take a look at something written in a different language on a different platform and they give me the deer in headlights look.

If only WotC would have taken my suggestion and partnered with developers of product that are already out there, proven to work they would have saved time and had a more robust product. I fear for the VTT, they have not had any experience in this realm compared particularly to the competitor products. And what about tools like Masterplan.

My best guess with performance has a lot to do with sub-optimal queries, lock contention, and performance issues with the database overall. We run a large website at my work, with very large databases and thousands of customers. When we first brought the latest version of the site online it was horribly slow. We bought the knowledge (read: Consultants) to fix the latency issue when our developers could not find anything further to tune.

 

[nerfherder]

My best guess with performance has a lot to do with sub-optimal queries, lock contention, and performance issues with the database overall. We run a large website at my work, with very large databases and thousands of customers. When we first brought the latest version of the site online it was horribly slow. We bought the knowledge (read: Consultants) to fix the latency issue when our developers could not find anything further to tune.

I used to specialise in database performance tuning, and your experience doesn't surprise me. I've seen otherwise clever developers select the full contents of a table into an array, sort in memory and then discard the unwanted results. I explained to him about WHERE and SORT clauses in his SQL, and that screen suddenly loaded up much quicker...

There is a huge difference between knowing how to use a tool, and knowing how best to use that tool (or even knowing which tool to use - I've pulled out critical pieces of PL/SQL and rewritten them in C because they did a lot of calculations).

 

[IronWolf]

If a software project is late (particularly a large one), it is often decided to release it prematurely, simply to get in some cash to be able to actually finish it. Promised features are cut from the initial release and the short respite gained is used to work on the first in a long succession of service packs.

Dropping some features to speed up a release is common and certainly helps one get some additional cash to finish what you first set out to do. Releasing extremely buggy software early is hard to justify in any scenario. You risk alienating customers with a buggy release hurting the whole reason you were releasing in a not complete state to begin with. It is a difficult balance to reach and it looks like WotC made the wrong call in this case.

An open beta would have served them so much better technically and PR wise if they didn't have the appropriate amount of internal staff for adequate testing. Folks that want to get a peak at the initial product, can file bug reports and even simulate load more accurately. All the while others can just avoid it until it is ready for mainstream.