h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

I was using the Pinnacle DC30+ capture plug-in for Adobe Premiere 5.1, but the discussion said this was a problem with all software that did this.

*UPDATE*
I apologise. I just Googled about a dozen sites to confirm this, then came across the following post in a BBS:

---------------------------------------------------------------------
start Regedit

search
HKEY_LOCAL_MACHINE\SOFTWARE\PinnacleSystems\miroVIDEOCapture\Settings

than add a new DWORD called:

No_2GB_Limit with the value of "1"
---------------------------------------------------------------------

It looks like Pinnacle put an artificial 2GB limit into that driver, so my "empirical tests" were meaningless. So, I need to plug the DV30+ back in and try a big capture again with this flag set... That won't happen this week (I just unplugged the card when I replaced my motherboard last night, and it's sitting in storage) but it will happen eventually, and I'll report back. In the interim, just assume I'm an @sshole and ignore me.

I just realised, that still doesn't explain how I was able to do a 4GB grab on FAT32...

but your example of notepad's behaviour in Windows XP was what I was replying to then."?

why were you replying to it's behaviour in XP when the question I asked specified WinME and Win98?

The place I get the idea they pretend it isn't complex is from all the FUD they put out. it all starts form the assumption that microsoft programs are both simple and intuitive, which is just untrue.

Further, their behaviour in the code is to make simplifying assumption (nothing wrong with that in principle), but they then either hard code them so that you can't modify such erronius assumptions, or hide them under obscure and unguessable parameter names in the registry.

You are right in saying that the registry has a search function, but it is almost unusable in most circumstances, and is not automatable in any case.

Also, the registry can only be backed up sensibly (to the extent it can be backed up at all) from regedit. Unfortunately this only works if the entire registry is uncoruppted, which it a condition it is notorious for becoming.

Am I suggesting having .ini files plastered all over the different drives? well, it certainly beats the idea of forcing all the stuff currently in mydocuments, mypictures and my music into a registry which has been shown reeatedly to be unstable and hard to scale up.

However I see no reason why they have to be splattered all over the drives. Unix does perfectly well having the /etc/ directory as a default location for them all to go into, with the option of having a directory with files in it if there needs to be more than one file for that program, and it makes things much easier.

Contrast this with the registry where instead of having all the information for one program in one place, it is usually scattered throughout the registry tree, meaning that most people have to get registry cleaning programs or reinstall every six months to 1 year, just because it doesn't do things properly.

as to other operating systems, I have used lots of unix derivatives (bsd, hurd, linux, solarix), the archimedes and amiga, cpm & mpm, and a host of eight bit and minicomputer operating systems, and none of them are as unstable and fragile as the 32bit offerings of microsoft.

Even under win3.11 you knew that the config files were usually either in c:\windows\system or in the directory containing the application, usually as .ini files where the text was fully accessable.

While I don't have a problem with the idea of a cetral location where everything is put, the registry proved itself incapable of being such a location back in win95, when it proved it's unreliability, fragility, lack of scalability and tendancy to rot in various ways. I have seen no significant improvements to it since, especially when the locations for parameters vary spectacularly from version to version of microsoft software.

When they introduced the monkey helper, and it proved a misserable failure, they removed it from the system. With the registry they seem to be bent on exacerbating the mistakes they have already build into it, and increasing their severity.

Considering the daily increasing likelyhood that they will go spectacularly bust, I am looking forward to it.

There are more advanced registry editors out there that include more advanced search but they cost money. There's also a registry repair utility included with newer versions of Windows.

I haven't had a registry corruption since I upgraded to Windows 2000.

I'm not firmly of an opinion one way or the other as far as the registry goes. It's a Microsoft tool, just like any other. Microsoft's tools mean Microsoft's rules.

I also think writing Microsoft off as going out of business is premature. Microsoft has proven its business accumen time and again. They're also not above unethical behaviour to advance their interests.

I think it is only a matter of time until they go spectacularly bust, but that does not mean that bill gates will go the same way.

Microsoft themselves have said that their profitiability is inversely proportional to the sucess of open source, which is growing it's market by 40% per year.

Also, microsoft has had to shift from selling you the cd with the software on it to selling you a yearly (per seat) license to use the software in order to keep profits fromdropping too fast.

Many people who don't currently use computers will not have an option of using a legal version of microsoft software due to a combination of increasing prices and local low wages. you just won't pay 3 months wages for a word processor when you can get open office which can do 95% of the same stuff for nothing.

Quite a lot of people won't use microsoft software due to the "microsoft tax" involved, either because they are poor, charities, or local authorities.

Microsoft are also being spectacularly unsucessfull in spreading their manopoly because too many people have seen other industries taken over by them.

Finally, microsoft are much more vulnerable to these problems due to exactly the same stuff which made their growth so spectacular in tbe first place.

They got a competetive advantage in the software industry by paying their programmers a pitance, plus generous share options. However the value of these share options is entirely dependent on the prifitability of microsoft, whose profits have been shrinking fast for years (hence the change in license conditions).

At some point, these share problems will make them worth much less, and at that point they either have to give even more generous share options (which is only an option in the short term) or have to start paying their programmers using the normal methods (which make them much more expensive, putting upthe cost of the software).

All these factors are working against the odds of their continuing existence, and are accellerating.

You think that the programmers' wages form a noticeable proportion of the cost of software? For any given retail box from a large corporation, the wages of middle management, marketing and other peoplr not directly involved in the coding itself have a much greater impact.

(peoplr = people )

It is not just the coders who get this deal at microsoft.

first, you have the coders.

then you have the armies of testers, debugging the code, checking it for usability, etc.

most of the cost of computers now is software.

lots of the cost of software is down to the number of people needed to get it working properly.

most of these people at microsoft get effectively minimum wage plus share options.

I don't see any way that their current model can survive their rapidly falling share prices and diminishing dividends on those shares.

It can't help but massively inflate costs at a time when microsoft are fighting desperately to extend their profit margin and find new monopolies to exploit (not very sucessfully in either case).

At the same time, there have not been any new "killer applications" for quite a few years, and computers are now cheap enough and powerfull enough that people ony have to upgrade to deal with feature creep incompatabilities.

As a lot of their current profits come from getting "box shifters" to install their software before the boxes are sold, and the software cost is forming an increasing percentage of the cost of a new computer, it can't help but have a serious impact on their profits.

after all, 700 pounds worth of software on a box with 1300 pounds worth of hardware is one thing, the same cost on 300 pounds worth of hardware is unrealistic.

Fair enough.


I predict that, in the face of reduced camera costs and increased computing power, the next "Killer Application" will be a low-cost Hollywood-standard desktop video editing package simple enough for Silver Surfers.

If so, it is likely to be on linux, which hollywood is already converting to en-masse.

However it has a fundamental problem that most other applications have not got.

it is relatively easy to build a photo gallery application, but when you move to video, you need all sorts of extra abilities in the editor so you don't end up with the sort of thing that looks like an amateur wedding video.

this is just something people don't know how to do, and unless you get it all correct, the result looks naff.

also, the cameras at the moment are largely a joke.

the sort of thing which would make it a killer application is using the massive memory of your camera to take tv quality video, but at the moment all of the video modes that I have seen on ordinary cameras are spectacularly daft.

The obvious example of this has to be the camera with 128MB of memory, letting you take 500*1200*1000 jpegs, or 50 minutes of video in 320,128?

a strange size, only used to give an attention grabbing number, and utterly useless in most circumstaces. better would be 640*480 and 10 minutes of usefull video in the same space.

Digital video cameras use even cheaper chips than the still cameras, but are still massively over-priced, and often have similarly attention grabbing useless features instead of more usefull ones.

one example I saw was a digital video camera which could record 1 hour of high res digital video, but could only export it over a usb1.1 connector, making it useless. the other option was to export is as analog tv signal, which lost the advantages of buying digital in the first place.

also, windows has various artificial limits bodged into it though their usual lack of forethought, which makes it likely that it will have to wait for longhorn before it is actually capable of dealing with the high volumes of data involved.

Well, remember that domestic VHS in the UK is equivalent to 352x288 resolution, 352x240 in the US. (Those are the standard resolutions for Video CD, designed to have a quality directly compatible to VHS...)

So, the camera you mentioned could probably store 30 minutes of VHS quality video, which is at least moderately useful.

it would be, it it had been configured that way, but the firmware is set up to handle one resolution only, and not very well at that.

also, you really need mpeg compression in the camera to do it justice, and the processors in ordinary cameras are vastly underpowered for that purpose.

also, these cameras tend to use proprietary formats with proprietary transfer tools, which often only work with one or two versions of windows, and then break.

also, they tend to have various error correction mechanisms locked on, and for genuine video work, you need to be able to turn them off.

If someone were to bring out a suitable, cheap camera which used open standards for the data, then it might form the basis for such development work, but nobody has done that yet.

Surely you downloaded the code, wrote a better installer and released that back to the open source community?

By the time I'd written a better installer, the versions would have incremented.