Good luck manipulation is one of the distinctive qualities of a TAS, one that unmistakably distinguishes a TAS from an unassisted run. This one manages to display that concept quite well.
If your movie didn't end after half a game, I'd actually vote yes. Maybe there's another incarnation of this game around, that's better suited for TASing?
interesting TAS. Despite today's date, I think it's an excellent TAS, even though it's longer than needed. Can't vote anything but yes.
nice to know you still got the time and the heart for TASing :)
No, Charlotte casts a spell in the same speed whether she is active or not (at least, for Sanctuary). Look at 020fcc1c if you want to verify it by yourself.
hmm, you're right. Looks like my brain is playing tricks on me.
Or maybe it was related to the sorceress crest, which wouldn't be feasible to get in an any% anyway.
at the beginning of the second part on nicovideo, you're doing some additional shopping, right after taking a detour to get more money. Looks like the detour can be avoided.
Purify can be cast faster when charlotte is the active character. It takes some luck to survive the cast and still have both sisters in effect range, but wouldn't it be faster?
I enjoyed the run, but IMHO the flaws are too obvious for publication. Are you planning to do an improved run someday?
There are separate Program Files directories for 32- and 64-bit programs. This isn't a problem as much as a "why do this?" thing.
because microsoft is proud to finally have workable 64bit-support and is determined to show it. Never mind that everyone else had that 7 years ago. ;)
Dacicus wrote:
yet I found Google Chrome, MS Works, and a 60-day trial of MS Office 2007 on this computer.
I speculate that the notebook vendor actually receives money (or other benefits) from microsoft for putting that office trial on it. If they just installed Open Office, microsoft would sell less office packages.
As a general rule, on each of my notebooks the first thing I did was to wipe the hard drive and manually reinstall windows and recent drivers. That reliably reduces the amount of crappy software packages on my system to.. well, one.
Not recommended unless you know what you're doing. Take care not to kill the recovery partition, if your system uses one instead of a DVD.
I am currently looking for nice legal uses for it though... I am sure I'll be posting quite a few.
any updates on that?
I still haven't figured out what happened in that smv (and how it should be useful when the game crashes afterwards), so I'm curious about actual uses.
Eye Of The Beholder wrote:
Just a curiosity about the arm pumpping technique:
How much time does it save in a run? Any% for example.
That's a difficult one, but I'll try ballpark figures.
Arm pumping increases your position by 1px/frame whenever used.
The ingame-time of the any% is 23.5 minutes. Subtract the time where Samus is rolling, jumping, damage boosting, shinesparking, fighting or waiting for something, and you'll end up with maybe 10 minutes of running (probably less).
Just to have some numbers, I'll estimate the average running speed to be 8px/frame, with arm pumping 9px/frame. Then we'd save 75 seconds just by running faster.
Add savings from shortcuts that are impossible without arm pumping. The torizo skip saves another ~20 seconds.
The TASers probably have better estimates, but I'm confident it's somewhere between 1 and 2 minutes ;)
not a bug, just a small suggestion for the polls: We currently have to click the radio button to vote. By the magic of <label>, you can extend the clickable area to the description as well.
The template snippet that generates this:
A good run on a sub-par game, yes vote :)
They tried to pull the "innovation" thing, improving the series in a few places (more varied level design) but totally screwing it over in others (invulnerability after hit, that stupid fumbly grapple, limited rolling, too many contra-isms..). Turrican 2 was better. </rant>
btw, the genesis version was the first to be designed, the amiga-version was the port (despite being released earlier). (source)
yay for more progress \o/
Could you explain what caused the repeating battle? Just a game bug when you didn't have any random encounters before, or is it something different / more complex?
yay, turrica.... wait, wtf?
good enough for a crippled version. Makes me wish once more for a working amiga emulator, but I fear that'll never be possible as long as .IPF is closed source and nondeterministic and .adf doesn't run anything but cracked versions :(
I would say that virtualization can take up quite a lot of CPU power. Nothing to scoff at.
I've done some heavy transcoding work on a windows-VM in virtualbox. I know two things:
- performance differed quite a bit depending on settings and virtualbox version.
- with the right settings, performance was very close to native windows.
As said, most of the performance problems in VMs are related to IO (disk activity or gfx). CPU power helps, but not as much as you'd like.
Sir VG wrote:
definitely try it out, if for nothing more then "how does the keyboard feel?"
Any new laptop keyboard will take a week to get used to. If only because the special keys are positioned differently.
There isn't too much overhead in virtualisation. The faster CPU will probably offset the virtualisation overhead in the most common scenarios. (Unless you're running IO or GFX-heavy applications, but no amount of CPU-power will make that faster).
Dacicus wrote:
They haven't been too sluggish for me so far (other than the start-up times),[..]
yes, mostly start-up times. Or opening a document or creating a new one.
I guess the larger amount of RAM in a new machine will speed that up anyway.
Dacicus wrote:
but someone brought up the point that this might wear out RAM faster than previous versions of Windows.
There's virtually no wearing with RAM. Of all the pieces of hardware inside your computer, the RAM is likely to live the longest.
Heavily utilizing RAM to improve performance is a good thing. The difficulty is to use it in a way that actually boosts performance, such algorithms often have to predict the future. If the prediction is accurate, performance goes up. If the prediction is wrong, the system wasted time.
Dacicus wrote:
You're saying that netbooks cost more just because they have a longer battery life (and lack a ton of near-essential features)?
no, because of the inherent difficulty of providing the same features within smaller space.
Netbooks aren't really more expensive though, they usually range between 200 and 400€, while notebooks are 300 to 1000€.
There is a third option: get an atom-based notebook. That'll get you a combination of a netbook's price and battery life with a notebooks larger screen.
In other words:
the decision between atom-based and high-end-cpu is a tradeoff between battery-life and performance
the decision between netbook and notebook is a tradeoff between portability and a larger screen
If it's meant to act as your main computer, I'd advise something with a large screen. In any case, try out a friend's netbook for a few hours before buying one yourself.
Based on your last post, I'd actually guess that an Atom-based notebook may be fast enough for now. It's faster than your old P4 in any case, and the only thing I see on your list that really requires cpu power is bochs. Gimp and OpenOffice may be a little sluggish (just like you know them), but they'll work.
On the other hand, an Atom may not be fast enough in 3-4 years, depending on your future needs.
So it's up to you to decide whether you want a lightweight device with long battery life or rather something with more muscle.
OS should be Win7 or Linux. There's no point in Vista and trying to stick with XP for another 5 years just won't work.
moozooh,
you correctly identified the problem that youtube + 720p + atom doesn't work well. I don't doubt that, but I object to the conclusion you're drawing: the CPU is too slow. It isn't.
Flash is just a crappy ill-maintained pile of mess with gaping security holes, which has done as much harm to web-security and web-standards as IE6. And it hogs CPU-cycles. When flash-games with the gfx and gameplay complexity of an NES game don't run at 60fps on a current computer, something is very very wrong.
Now there's several possible solutions:
1) watch at 480p
2) use a different method of watching youtube-content. Being a bash-monkey, I use youtube-dl + mplayer. There are a multitude of firefox addons making this even easier.
3) blame the CPU, then spend more bucks on a bigger CPU you don't actually need, costing money, battery life and adding weight.
The reason I objected is that you're going for 3) without even considering a combination of 1) and 2). Completely voiding the advantages of an atom-platform for a single, non-essential use-case that can be worked around is a bad move in my book.
oh, and just for the record: I love javascript. I have a job developing browser-based games, using php/mysql on the server and nothing but sweet html/javascript on the client.
We know about flash, but whenever we decided to use it in a project it has bitten us - narrowing down target platforms, introducing funny bugs across versions, generally not working for a good portion of our players (hello support costs!) as well as being a pain to develop in a team, to deploy incrementally etc etc
oh, I'd love to have an atom-based, fanless netbook instead of a notebook. It's plenty for everything I'd use a portable computer anyway.
But if the notebook has to double as the main desktop computer (like Dacicus wants), it's probably a poor choice.
720p youtube is a poor argument though. Why would you buy a faster, hotter, more expensive CPU just to compensate for a totally crappy piece of software called flash? Even youtube offers a <video> based alternative now.
while waiting for the blue warp after odolwa to get activated, I did some weird playaround, which confused a lot of people, even a lot of experienced players ;)
It's probably a good idea to show what you're waiting for. Like keeping the shopkeeper visible while doing stunts or (unsuccessfully) runing through the unactivated warp before showing off. Or, if you're waiting 10s for the warp, you could do some stunts for 9s, then stand in the warp idly for the last second, just to show "look, it doesn't activate yet". [1]
It's technically much more precise to reach the warp at the earliest possible frame where it activates - but that can be confusing for the viewer. Precision is worthless if it actually makes the run look sloppy.
Of course this only applies to longer waiting periods. Smaller ones, like the bomb shop, are a matter of opinion. I don't mind 1-2 second downtimes at all, there's little chance to blink otherwise [2]. I'd say: if you can do something fun without being confusing, go for it. Otherwise please remember that the thing that looked sooo cool in frame advance will be over quite a bit faster when watched at full-speed. Don't overdo it, sometimes it's better to be idle and give the viewer a chance to keep up.
But you'll be the judge of that, considering your previous work I'm sure you'll do awesome :)
[1] Disclaimer: I haven't watched the old TAS in a while and don't remember what you actually did
[2] if you ignore the cutscenes.
Lenovo Thinkpads are durable and they will probably last for those 6-7 years you're thinking of.
IBM thinkpads sure were, but the stuff Lenovo sells seems on par with all the other cheap consumer trash out there. My R61 is 1.5 years old and the CPU fan is basically dead. The whole thing doesn't run above 800MHz for more than a few minutes and makes noise like a woodchipper.
We bought three of those, one for me, one for my brother and one for my girlfriend. They don't use their laptops as frequently as I do, but theirs too show signs of degradation.
In other words, don't expect a notebook to be any better just because there's a "Lenovo" sticker on it. Unless you like Track Points, which I (unfortunately) haven't found on any other brand yet.
I wanted the 'optional parameters' to be as explicitly addressed as possible.
because writing FrameControl::Priority is more explicit than 'Priority'? I'd argue that the explicit version is even worse, as the scope is apparent by context and you're just adding noise.
Unfortunately the consts don't give you compile-time-checking either (meaning: typing errors aren't caught by php -l), so the only reason to prefer the consts would be using a smart editor with auto-completion and const-highlighting.
(btw: if you know such an editor that works on linux and doesn't take 3 minutes to start like eclipse, please tell me ;))
Hoe wrote:
There's also a useful quirk to the constants, you can have 'aliases.' Like:
const Property = "blah";
const Alias = "blah";
add
class MyClass extends Options {
...
public $aliases = array(
'Alias' => 'Property',
'Alias2' => 'AnotherProperty',
);
}
and you can hide the corresponding logic for alias handling inside the Options class. Not a reason to use consts.
I think that can be done much easier without relying on those static const members. consts are a pain in php, as they cannot be overloaded. As you noticed, your implementation relies on internal class specifics that may change between versions.
Here's my design:
class MyClass extends Options {
public $options = array(
'optionname' => 'default value',
'secondoption' => null, // no default
'thirdoption' => 'third default',
);
public function __construct($initialvalues = null) {
parent::__construct($initialvalues);
}
}
I didn't bother writing down the respective implementation of Options, but it should be shorter than yours.
Any use cases my design doesn't cover?
(Can you remove the old list to see how just the new list would look?)
No, that's technically not possible. You can however make a simple screenshot and remove that area yourself. :p
I added a third option, with headers as suggested. The format column "(fm2,mp4)" was killed, too. I dropped the game-suffix "(USA v1.0)" as well (IIRC not supported by the database, but a simple preg_replace('/ \(.*\)$/', '', $gamename); should do the trick).
/edit: right-aligned the time column while I'm at it.
