What is Hurting WoW’s Polish?

Warning – Fanboi’s may attack!

WoW has changed. Change is sometimes a good thing, and I’m not going to belabor the obvious changes (less challenge) or industry misconceptions (subs are needed to cover ongoing costs) instead, I am going to look at two things I have noticed the most since my foray back into WoW.

First, to stave of the flames – WoW is still a fun game. It presents well, plays well, and yes, we all know how many subs they have. They are obviously growing still, but there are two nagging things that jump out at me that aren’t typical Blizzard “style”, every time I log in and (try) to play. Confusing conundrums, after the cut.

1.) “All Instances are Full – Please Try Again Later”. Wait.. what? The company that has the largest player base, highest gross revenues, AND “gigantic” profits can’t afford enough instance servers? C’mon. That is downright silly irresponsible. You overcharge subscription fees under the guise of “ongoing maintenance costs”, yet don’t use those revenues for, uh, maintenance costs? Give me a break. This is inexcusable. Granted, sometimes it is only 2 minutes. Two nights ago it was 20 minutes. [waits for someone to come in and say “duh, dude, it was only 20 minutes, go get a sandwich or something” – that’s not the point. It’s a point of principle.]

2.) “We are performing Rolling Restarts on all/x/y/xyz realms”. The majority of the time I log in this message is displayed on the login screen. I don’t log in every day mind you, but my last 7/10 logins I was greeted with this message – some/all/few servers require rolling restarts. See (1) for “ongoing maintenance costs” as a big question, but these are Blizzard sanctioned servers, written in Blizzard code.  Are they out of your control now? Or is it something else?

What’s up, Blizzard?

Possible issues:

Code is too big, too convoluted, too confusing. I can only imagine the tangled web of code for WoW, with three expansions under the belt. Do things get lost in translation, from one coding team to the next? Is it not so neat and tidy as it once was? Too much going on? I always wonder if the programmers think about “how they would structure the code if they new back then what they knew now”. I can imagine the code is a mess – I could be mistaken, but just the thought of what is required for a game of this magnitude, and the various levels and personalities of programmers who have been involved over the last fiver years altering it. Someone with programming experience can let me know if I am far off on this one. I can just imagine.

Return on Investment trumps quality in the new Blizzard. I am leaning towards this one. You know they can afford new servers (whether or not the code can handle them is another question altogether). You know the cash is there – unless, of course, it is more important to put that cash into dividends instead of instance (or stable) servers. Instead of investing in staff for multiple specialty jobs (improvements, patches, stability) keep the same small teams that work on everything – one at a time. ROI, baby. Unfortunately, not ROY.

Perhaps it is a combination of both. Regardless, anyone who has been playing the game lately can’t argue that the game seems to be less smooth than “back in the good old days” with nagging inefficiencies and error messages – that aren’t being solved. 

I was going to cancel again, but I really think they need my $15 a month for “ongoing maintenance”. Quitting now (again) will just hurt the rest of you folk who live by WoW – and I wouldn’t do that to you.

15 comments / Add your comment below

  1. Hmm! Well, I don’t have time for WoW, but here’s an interesting bit: Runes of Magic has been getting _more_ polished with every patch. I keep thinking “wow, this is the sort of quality I would expect from something like Blizzard”.

    The interface has become smooth, servers were never down when I tried to log on and there are always a bazillion players zooming by wherever you go.

    Have you tried it yet? I think they have a North American operator as well, so it should be fine from Canada.

    Some translations are a bit buggy still (and I swear I’ve seen some Chinese characters in an option screen once), but I’m really surprised, especially for a F2P. It’s not a WoW clone like many people say, though, that’s only superficial. So if you expect a WoW clone when you get into the game, you might be disappointed.

  2. When I started working on Meridian 59, a bit over a year after launch, the game had already become a really complicated. Four different development teams had been working on the old girl by that point, and they had all left their impressions on the game. The game was also doing things it was never originally designed to do. Keep in mind that this is a game where a shard runs on a single machine, so we’re not talking about any black arts of distributed computing yet, either. Just cruft built on cruft through many different hands.

    It is four years later and the gods only know how many hands have passed over the code by now, so it’s starting to show it’s age, I’m sure. Every time there’s significant downtime, I tell my friends that, “Somewhere, a programmer is crying.” I know that from personal experience, because I was that crying programmer for several years when it came to M59.

    This isn’t to say that some squirrelly business decisions haven’t been made. After bragging about making a cool billion on the game, you’d think that they could dedicate some effort to shoring up the leads in the dike. M59 doesn’t run perfectly even now, but our tiny team making below minimum wage was able to get the game running fairly stable, to the point it runs with only one person making sure the servers don’t catch on fire. (I take no responsibility for the community, though. :P)

    Some thoughts from a developer.

  3. 1) Totally agree. And, apparently, it’s been going on long enough that Blizzard should have done something about it. I haven’t done my due diligence, any word on the official forums about it? Are people complaining?

    2) I’ve been logging in almost daily over the last week, and haven’t noticed the rolling restart messages. Back in the winter when I did see them, though, it was mostly in the early mornings. Of course, that’s when I was actually playing, and not really logging in during prime times… Plus, I don’t pay super-close attention to those messages on the sidebar when logging in. Perhaps I should

    Though, I’m not sure that “polish” is the word I’d use to describe the situation… As I’ve been going through Northrend content, I’ve found it to be more polished than almost anything in Burning Crusade. Though, I suppose we may be talking about different things, but other than the instance server thing, WotLK doesn’t give me the impression of a world that’s pretty on the surface, but breaking down behind the scenes…

  4. I’m an IT Specialist, and we have regular restarts of the servers in our farm usually once a month. Why? Patching. You can have the most stable bit of code in the world – we only run read only websites for the most part – but if you anti-virus and operating system patching needs to be done that tends to need a restart of the server like it or not. You also have to do semi-regular disaster recovery tests on bigger server farms if you’re doing things properly, so any of these things can be attributed to why the servers are sometimes down.

    HOWEVER – with that much money you can have mirrored servers for every realm meaning that you can restart things without EVER having to take users offline. You can also afford to stick another 10-12 blades (yeah I work for IBM) into the rack to allow more instances at the cost of a drop in the ocean compared to what Blizzard raked in in the last few days let alone months.

    So yeah – if money was na issue then you do still need to restart things from time to time. When money isn’t (which it shouldn’t be in this case) then there’s just no excuse.

    @Psychochild – I’ve also been that crying developer more times than I can tell you (though not in the gaming world) – its made worse as well when the customer has your mobile number and can shout at you!

  5. @Ramon: I do want to give RoM a whirl, I am just so pinched for time. I have M59 slated for my next new game tryout. It’s on the list though. Blizzard was good at that in the past, continued updates, things getting better – which is why I guessed the code has reached critical mass.

    @Brian: Thanks for the inside look – it was pretty much what I was thinking, and it doesn’t surprise me one bit. I can just imagine the difference in code size and complexity from M59 to WoW. While I would guess they would have had made some good organizational choices in the beginning, who knows what it looks like now.

    @Pope: The game itself is polished, but my argument is that constant rolling restarts and/or unable to access game content is a lack of “polish” on it’s own. I think it is a fair argument. It doesn’t matter how much your gameplay is “polished” if your customers can’t experience it. Your pretty on the surface, breaking down underneath is the gist of what I am getting at – very much agree.

  6. I did the WoTLK trial, was fairly impressed and resubbed for a little while. I do play at odd hours, certainly not during prime time. That said, I have to agree with Chris. The amount of server downtime (scheduled weekly downtime, frequent restarts) was striking. I don’t see how it could entirely be a “multiple teams” issue. There are much older MMOs that don’t have nearly the downtime that WoW seems to.

  7. BFG50: You need to restart your servers? What OS are those running?

    We usually restart ours when a security-relevant kernel upgrade comes along, which is after 100 – 1000 days. My personal uptime record is somewhere around 1200 days. The machines are quite busy, we run several hundred websites and a bunch of dozen applications.

    These days they’re on about 30 virtual machines running on two hardware nodes, but it used to be a real physical server for every virtual one back in the days. We had more outages back then due to hardware failure. Now the hardware nodes have redundant everything except CPUs, because to afford that, we’d have to be Sun customers, and we’re too poor for that.

    Not to flame or anything, but the only reason you should have to reboot a properly designed OS is hardware failure or a kernel upgrade, and they’re working on hot-swappable kernels these days.

  8. This is WAY off topic but Meridian 59? Holy crap I started playing the Alpha/Beta of that game back in, oh 1994-1995? When they released the game I had been playing for a year and loved it, but a decade and a half ago I couldn’t imagine paying the 10$ a month or whatever it was.

    It both blows my mind and makes me happy to know the M59 is still around in some form or another.

  9. @BFG50/Ramon: All of your smart tech guy talk is hurting my head. Is that kind of server structure available in a WYSIWYG format? (grin)

    @Aaron: It’s still up and running, and I like to “vote with my wallet” so to speak and support small devs/indy teams who support good design decisions. Besides, I never played it in the old days, so will be a brand new (VGA?) experience for me!

    Okay back on track. Uh, Blizzard, buy more instance servers or something.

  10. Ok, to respond to the first point (finally?)… since 3.2 went live NO ONE has been able to get into the Stockades in Stormwind. One fellow guildie spent 2hrs trying to get in the other night. OMG!! C’mon, Blizzard… get your shite together and get the servers going. If the battlegroup instance server is getting crowded, then make an instance server for each bloody realm. This is getting effin’ retarded…

    /end rant

  11. I think it is perhaps they dedicate more resources to the newer instances (which would make a tiny bit of sense) – the problem is, of course, if your servers are running at an average of 100+% capacity shifting around the instance priorities is going to have the effect above. Nothing short of adding new servers (or code, if that’s the problem) will solve the issue.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.