Recent Posts

Pages: prev • 1 2 [3] 4 5 6 • next

The Form Letter Machine / Re: Request for var file-list precedence behavior for variable overloading

« Last post by superticker on October 01, 2007, 01:19 AM »

... one thing you hint at ... is that this will be much more oriented to people who are managing largish size collections of forms and data....
-mouser (September 30, 2007, 02:23 PM)

Actually, I believe it's to one's advantage to have as few form templates as possible; otherwise, you're repeating (and maintaining) like information between templates. I have three templates each with over 60 check boxes and radio buttons. Perhaps I can send you a sample tree.

I also find the current design reasonable for my work. The few suggestions I do have were discussed in the bugtracker a year ago, and they're simple mods.

Another application--totally separate from TheFormLetterMachine--would have an SQL database generate output files based on stored database fields. I have a real need for an application that does this as well, but this is a totally different app which should have a different name (such as MS ACCESS or MS Excel). That's really not what TheFormLetterMachine is about though.

If someone could recommend a flat-file database that can selectively convert fielded records into a formatted output file, I would really appreciate it. I produce a monthly news digest of ballroom dance events. Each record would be one event with fields like date, time, place, price, location. I need to select the records relevant for the given month, and produce an output file that forms that month's digest of events. I thought about using MS Excel's output file capability for this, but there's probably a better way using a generalized flat-file database. Can someone please suggest one?

If the flat-file database can produce an XML output file (digest) that can be posted on the web (monthly newsletter), that would be even better. In fact, if it can produce an XML output-file schema compatible with an RSS 2.0 feed, that would be perfect. Then I'll just publish my monthly newsletter as an RSS feed instead. But this is really a database problem, and not a problem for TheFormLetterMachine to attempt (which is a different application).

The Form Letter Machine / Request for var file-list precedence behavior for variable overloading

« Last post by superticker on September 30, 2007, 02:18 PM »

Currently, TheFormLetterMachine only allows one symbolic variable file at a time and you must choose between them. It would be significantly better if TheFormLetterMachine operated much like a linker such that multiple object libraries are opened at once and the precedence of the library listing determines which values the overloaded symbolic variables would take. For example, consider overloading the printf() symbol in debug.obj:

$ LINK main.obj,debug.obj,subroutines.obj > main.exe

would produce a main.exe load module (executable) with additional printf() debugging behavior for tracing the program's progress. In contrast, if we reversed the debug.obj and subroutines.obj list order, ...

$ LINK main.obj,subroutines.obj,debug.obj > main.exe

then list-order precedence would resolve the overloaded symbols by the subroutines.obj library first instead of the debug.obj library, so the main.exe result would have less debugging objects linked in with it (although non-overloaded symbols within debug.obj would still be included in main.exe).

To use an example more appropriate for TheFormLetterMachine, say a computer scientist is looking for a job in either the graphics, database, or networking fields. When a recruiter responds to his job ads, he would respond with a precedence listings:

database.var,graphics.var,network.var,general.var --> for database job inquiries
network.var,graphics.var,database.var,general.var --> for networking job inquiries

This means that the precedence list (of overloaded symbols) will need to be juggled for each e-mail response, so there needs to be a "quickie" drag-and-drop GUI to juggle the precedence list each time. In addition, the command line should accept a var-file list, such that a Windows shortcut can define a given precedence listing.

My only reservations with this approach is that some end users may not understand the concept of "precedence" and "overloading", although they could still use TheFormLetterMachine the "old" way without this understanding. However, if they wanted to become power users, then they would need to understand these concepts.

The Form Letter Machine / Re: To save or not to save upon quitting; discarding temporary mods

« Last post by superticker on September 06, 2007, 12:45 AM »

There's been some talk of re-organizing the form letter machine user interface in a more dramatic way, which i think is probably needed, and i think we should revisit these issues then.
-mouser (September 05, 2007, 10:56 PM)

Well, this fix isn't that critical. Perhaps your time would be better spent on re-organizing the GUI instead. So, to start a new thread, what's wrong with the current GUI and how is the new GUI design going to be better?

The Form Letter Machine / Re: To save or not to save upon quitting; discarding temporary mods

« Last post by superticker on September 05, 2007, 10:28 PM »

... When you change a tree node text item from the main screen, it assumes this is something you just want to change temporarily. i.e. it *assumes* that by changing it here and not in tree configuration screen that you do *not* want to save it, that's why it doesn't prompt you....

One way to solve the problem is add an option about whether changes made on main screen should be treated as changes to the tree that presumably should be saved (and thus user will be prompted if they try to change tree or exit without saving them).
-mouser (September 05, 2007, 05:07 PM)

I'm trying to decide if adding another configuration option is more or less confusing than simply providing two Quit choices: (1) Save and Quit, or (2) Quit w/o Saving. Somehow I think the latter approach is more intuitive than adding another config option.

The Form Letter Machine / To save or not to save upon quitting; discarding temporary mods

« Last post by superticker on September 05, 2007, 11:20 AM »

... [To create a temporary comment in a note] just design your configuration tree as normal and save it, then when composing a letter, select the node you want to edit, and edit it in the top box above. Changes in this top box are just temporary and last for the duration of the current letter only.
-mouser (July 05, 2007, 11:30 AM)

Now I understand why there isn't an automatic save feature on The Form Letter Machine when quitting. In order to give the user the choice of (1) saving or (2) not saving upon quitting the program, I would create two mutually exclusive Quit choices: (1) Save & Quit and (2) Quit w/o Saving.

With the single Quit choice, I'm constantly loosing my changes because I forgot to save them before quitting.

The Form Letter Machine / Auto saving tree and variable files

« Last post by superticker on June 16, 2007, 12:48 PM »

... the protection feature that warns you that you're closing the app without saving the tree sometimes forgets to warn you.

let me fix this, and then decide whether auto saving tree changes should be automatic.
-mouser (June 16, 2007, 12:09 PM)

In practice, one does more saving than aborting of changes, so in this sense, auto saving trees and variables on program exit makes sense. I have an accounting program (Simply Accounting by ACTPAC, excellent product) that works by auto saving. At first I was concerned, but as long as the File > Revert option is available, it hasn't been a problem, and I have never lost a thing. I rarely even open the File menu on my accounting program.

There is an issue that when you switch trees, you'll need to auto save on the switch. That means if you switch back to the original tree and want to revert, you can't (unless the program saves *.bak files for trees). However, I'm using several tree files and I've never experienced a case where I wanted to do that. And with the present implementation, doing this maneuver would still be a problem.

The Form Letter Machine / Re: Making command line arguments work

« Last post by superticker on June 16, 2007, 11:38 AM »

Can you try this new build:
https://www.donation...ormLetterMachine.exe

(It seems that in the last version ... the -out=stdout option [was being forced])
-mouser (June 15, 2007, 01:23 PM)

I just tried it and it works. But there's a new issue. If I close the app dialog box (instead of using the [close] button), TFLM writes a blank output file. I don't think any app should ever write a blank file. I would suggest:

If the application window is closed, treat that as pushing the [cancel] button so that nothing is written--not even a blank file.
Change the name of the [close] button to [Output & close]. Honestly, I didn't even understand the real function of the [close] button until I read this forum.

I might also add that the protection feature that warns you that you're closing the app without saving the tree sometimes forgets to warn you.

I'm wondering if the save should just be automatic on program exit, in which case the "File > Save tree" menu item should be replaced by a "File > Revert changes" to restore (reload w/o saving) the previous tree?

The Form Letter Machine / Changing the default data (tree) directory

« Last post by superticker on June 15, 2007, 03:23 PM »

This is an old issue brought up before in the Bug Tracker, but here it goes again.

If on the Start In: field of TFLM shortcut, I redirect the data files to be stored in

[Select]

%HOMEDRIVE%%HOMEPATH%\TheFormLetterMachine

directory, the application ignores that and still puts them into the %ProgramFiles%\TheFormLetterMachine directory just the same.

Honestly, for Windows 2K, XP, & Vista compatibility, the app should be putting them in the %APPDATA%\TheFormLetterMachine directory by default, which would work for me. The Start In: field still should be able to override this default. (I thought the MS foundation class libraries behaved this way by default, but I may be wrong.)

On an unrelated note, I'm bring this up now because I had to recently lock down my machine(s) to prevent innocent-looking spam URLs from indirectly linking to exe files that try to download and install on my system. Clearly, this is a new trick by the spammers to install back doors on machines. But I now have to grant access exceptions to directories like %ProgramFiles%\TheFormLetterMachine and "%ProgramFiles%\Clipboard Help+Spell" to circumvent this security to get these programs to still work.

The Form Letter Machine / Re: Making command line arguments work

« Last post by superticker on June 15, 2007, 11:08 AM »

... try dropping the double quotes altogether from the -out parameter. Since there are no spaces in the filename, they shouldn't be required.
-Boxer Software (June 15, 2007, 10:45 AM)

Yes, but there's a colon in "C:\" that needs the double quotes. I tried getting around that with symbolic variables (see earlier reply), but that hasn't worked either.

The Form Letter Machine / Re: Making command line arguments work

« Last post by superticker on June 15, 2007, 11:04 AM »

My guess would be that it should read:
[Select]
"C:\Program Files\TheFormLetterMachine\TheFormLetterMachine.exe" -out="C:\users\mehl\announcement.07.txt"
I moved the " from before the -out to before the C:\ at the end.
-cthorpe (June 15, 2007, 10:29 AM)

I agree. Thanks for the correction. But it still doesn't work. I also tried removing the double quotes altogether using symbolic variables:

[Select]

%ProgramFiles%\TheFormLetterMachine\TheFormLetterMachine.exe -out=%HOMEDRIVE%%HOMEPATH%\announcement.07.txt

But that's not working either. Strange.

The Form Letter Machine / Making command line arguments work

« Last post by superticker on June 15, 2007, 08:58 AM »

I setup a Windows shortcut with the following command line parameters:

[Select]

"C:\Program Files\TheFormLetterMachine\TheFormLetterMachine.exe" "-out=C:\users\mehl\announcement.07.txt"

However, when I press the [close] button on The Form Letter Machine, the anticipated announcement.07.txt output file is never generated. What am I doing wrong? I'm running v1.04.01.

Clipboard Help+Spell / Re: Memory Usage

« Last post by superticker on June 14, 2007, 12:20 PM »

I will add a feature i add to most of my programs that let's you choose whether you want them to release their working memory when minimized.
-mouser (May 04, 2007, 07:55 AM)

I think it's standard practice to have Windows apps release their memory (like for their GUI) in the free-memory pool when they're minimized and the zero-page pool if they never need that memory again. Understand, the first pool is for temporarily releasing memory that otherwise needs to be reclaimed when restoring the app to the desktop. The idea is that if no other programs need that space, then those pages remain available to the app that originally released them when it gets restored; (otherwise, those free pages can made available to external apps instead of calling the swapper).

The zero page pool is for permanently releasing memory (never to be reclaimed in the future). The OS immediately writes zeros in these pages (to prevent browsing) so they can be immediately allocated to unrelated processes.

The Form Letter Machine / Re: Variable Management (Requests)

« Last post by superticker on June 14, 2007, 10:14 AM »

4 modifications / features ... for configuration trees with many variables.
Only display the variables used in the currently activated portions of the tree. For example, if my currently selected check box items and radio button items make no use of the %MeetingDate% variable, the variable should be hidden.
-the3seashells (April 09, 2007, 03:29 PM)

This is a good feature, but there needs to be some way to disable it so you can locate discontinued variables and remove them.

Allow variables to be clicked on and filled in as in a form in a word document. When looking at the entire document (the bottom left area where the entire letter is shown), it would be much easier to type in the information there than hunt for, and then update the variable from the variable list.

Clearly mark unfilled variables in the ‘completed letter preview’ area. I find myself spending the majority of my post TFLM proofing time checking to make sure I have not forgotten to included / replace all of the needed variable names.
-the3seashells (April 09, 2007, 03:29 PM)

Well, there are really two distinct types of variables: (1) predefined variables and (2) fill-in variables. They both need to be handled differently. The TFLM currently handles the first type well.

Firefox handles fill-in variables by making suggestions as you type in the fill-in field. For filling in my name and address on web forms, I like Firefox's method. Firefox also has a method of deleting obsolete fill-in suggestions, but it requires a cryptic hotkey combination I can never remember. I also like this delete-suggestion feature, but I wish it was done with a right-click context-menu item instead of a hotkey combination.

Living Room / Re: adding SATA to a non-SATA motherboard

« Last post by superticker on December 23, 2006, 12:58 PM »

if you have a 5yr old motherboard, perhaps the answer is to buy a new motherboard (or new computer?). I realize that's easy to say and not so easy to afford, but at some point trying to upgrade these things is more trouble than its worth..
-mouser (December 20, 2006, 12:52 PM)
Argh, I think you're right.
-superboyac (December 20, 2006, 05:35 PM)

I agree. If you're going to buy a new disk, I would indeed get a SATA drive. But in addition to that, I would replace my motherboard with one which had a SATA interface as well as a PCI Express graphics interface. That means buying a new PCI Express graphics adaptor that's compatible with Vista. You can still keep Windows XP, although if you plan to upgrade to Vista, I would do so when you replace the disk drive.

The only thing that is important and lacking from my system right now is a robust backup method, which is why I'm splurging on the two 500GB hard drives.
-superboyac (December 20, 2006, 05:35 PM)

Now that's a whole different issue. If you're looking for a backup disk, I would buy an external drive with an "eSATA" interface. This will have a little different connector than the standard "internal" SATA connector to facilitate better shielding for longer cable length.

When buying an external drive, if you're a computer professional, I wouldn't fool around. Get something like the G-Drive, which has everything: Aluminum case, no fan, eSATA, 400 & 800MHz Firewire, and USB 2.0. Of course, the eSATA connection will be the fastest followed by the Firewire connections. Here's an example source: http://www.academicsuperstore.com/market/marketdisp.html?PartNo=752728

If you're not a computer professional, getting an external drive with only an eSATA and USB interface would be good enough. If it's a high performance disk (e.g. it gets hot), I would still shoot for the aluminum case w/o the internal fan. Slow (cool) drives can be put in plastic cases.

The G-Drive I mentioned above allows air to circulate both above and below it so it's cooled on both sides. That's what you want to shop for if you're buying a high performace (runs hot) drive. Most people may be happier with a slower, cooler-running drive though.

General Software Discussion / Re: Warning: FireFox renders slow like a turtle in a coma

« Last post by superticker on December 15, 2006, 05:40 PM »

Must be some extensions playing tricks, mouser. FireFox rendering does seem a slightly tiny bit slower than IE, but not much,
-f0dder (December 14, 2006, 09:35 AM)

I agree. My FF 1.5.0.8 is very fast. You never told us whether this slowness was due to excess page faults (most likely!), CPU time, I/O waits, or network slowness. Can you fire up Procmon (Process Monitor from System Internals) and give us a screen shot of the Performance tab results for Firefox? (How do you post a screen shot on this forum anyway? I would post my Performance tab, but I don't know how.)

For my memory working set (on Firefox), I'm seeing 81,000K. Yours will be somewhat higher if you have some extensions that hog memory. If your working set is really large, then tell us more about your Page Fault delta (which will cause I/O waits).

You may be able to tweak some parameters in Firefox to reduce Page Faults, but you'll probably need to buy more memory chips when you do. Uninstalling excess extensions may be better alternative.

Mini-Reviews by Members / Re: Mini-Review of Fineprint (Virtual Printer)

« Last post by superticker on December 13, 2006, 10:41 PM »

Concerning Fineprint, is it possible to configure it to print 12-up as well? I'm asking because I sometimes design tickets for events (like dances), and I want to print 12-up and 16-up on a page.

Does anyone remember DynoPage on the Mac? One nice thing about DynoPage is that you could design your own layouts and margins for these different layouts. This let you use non-standard paper sizes and these new paper sizes would show up in your Page Setup screen for any particular program like PageMaker. DynoPage didn't have a 16-up "tickets" layout, so I designed one for it complete with cut marks for dicing the tickets.

DynoPage was also smart. It did much of its manipulation within the PostScript interpreter of the printer. Of course, you would have to be using a PostScript printer to take advantage on this. As an example, you can page lock a watermark in a PostScript level 2 (or 3) printer so it prints out on every page automatically.

I often thought about learning PostScript but I really don't have time to learn yet another programming language. I wish I did know it though.

Developer's Corner / Web development tools that integrate with Firefox

« Last post by superticker on December 10, 2006, 03:25 PM »

Here's a website that lists all the great web developer tools, including FireBug and the Web Developer Toolbar. This site is also discussed in my StumbleUpon blog (linked on my DonationCoder profile).

http://cyber-knowledge.net/blog/2006/10/05/20-firefox-extensions-that-every-web-designer-should-know-about/

General Software Discussion / Re: NTFS encryption security on portable drives

« Last post by superticker on December 03, 2006, 02:34 PM »

I have strongly discouraged users from formatting their USB flash drives with NTFS directories if they are taking them outside their Windows domain for fear it might create ownership problems down the road.
-superticker

Not a good suggestion though - if somebody formats his drive as FAT, he'll be in a nasty situation once he's filled up some 100GB and need a file that's >4GB large (but okay, at least there's transparent conversion to NTFS with "convert.exe").
-f0dder (December 03, 2006, 02:17 PM)

Actually, all USB flash disks (and ZIP disks) are shipped formatted as FAT. SanDisk uses FAT16, and that would limit the flash volume to 4 GBytes. I guess I don't know why SanDisk doesn't use FAT32. Does anyone know?

The problem is that users want to convert their new USB flash drive from FAT to NTFS, and that's what I'm discouraging until I figure out how foreign NTFS file ownership would be handled between office (central domain controller) and home (foreign host).

General Software Discussion / Re: NTFS encryption security on portable drives

« Last post by superticker on December 03, 2006, 12:41 PM »

Windows stores the EFS encryption key, encrypted, in the registry... for domain logons, I assume it's stored on the domain controller. For non-domain machines, you'll probably need to make sure that all machines have the right credentials, and perhaps SIDs as well.
-f0dder (December 03, 2006, 11:14 AM)

The problem with NTFS files is they have the concept of "ownership" attached to them. If that ownership is attached by a central authority (domain controller), then switching disks among domain members shouldn't be a problem. But when you mount a "foreign volume" from outside the central authority, then who owns these files? ...the Default User?

Should you even be able to mount a foreign volume? If so, then who takes ownership of the Default User's files? In this weird case, I "think" the Default User would be the local administrator since the creator of the original domain account to which these lost files once belonged to would not be available on a foreign, non-member host. The other possibility is that there is no defined Default User; therefore, you can't mount the foreign volume.

I have strongly discouraged users from formatting their USB flash drives with NTFS directories if they are taking them outside their Windows domain for fear it might create ownership problems down the road. Even if those flash drive files are owned by the Everyone group, it's still the Everyone group for that specific domain, not the entire Windows world.

If there is a safe approach for defining NTFS ownership on portable (foreign) disk volumes, could someone step forward and explain this? For security reasons, I don't like users using FAT volumes, but for portable disks, I'm not sure how to get NTFS ownership to work.

General Software Discussion / NTFS encryption security on portable drives

« Last post by superticker on December 03, 2006, 01:58 AM »

If you use NTFS on the portable drive, you can use Windows' built-in EFS encryption.... Doesn't work on "Home" editions of XP, though.
-f0dder (April 21, 2006, 08:23 AM)

This comment brings up a question I have about NTFS file security. If I move an NTFS disk between two Windows Pro machines belonging to the same domain (and using the same enterprise license key for Windows Pro), the encrypted files should be okay (if they're authenticated with the same domain controllers), right?

What if I move an NTFS disk to another Windows Pro system that's part of a different authority (different domain or difference license key)? Won't--or shouldn't--those encrypted files be unreadable? Or am I missing something here?

Will you even be able to mount an NTFS volume that comes from a foreign domain (or license key)? My understanding is that foreign NTFS volumes present mounting problems, especially when they don't have Everyone read/write access. Does someone know a reference that discusses this more?

Some backup software (like Paragon) lets you change the volume SID on an NTFS disk, but I always thought you had to decrypt all files before doing so or bad things would happen.

General Software Discussion / Re: Windows memory-paging behavior

« Last post by superticker on November 21, 2006, 11:40 AM »

This helped me understand a bit more about windows memory handling: http://shsc.info/WindowsMemoryManagement
-nontroppo (November 21, 2006, 07:40 AM)

That's an excellent description about the Windows memory-paging system. The important thing I learned is that an application's active virtual working-set (of memory) is automatically reduced when that application is minimized. Moreover, its unneeded virtual pages are returned to the "standby" page pool where they remain intact but are up for grabs. It's much like moving a file into the trash. It's still there (but up for grabs if needed). When the minimized application is restored, those same intact standby pages are placed back into service without creating a page fault to disk. Very nifty.

This also explains why applications page while idle when there's plenty of memory. They're simply returning pages they won't need for a while to the standby page pool just in case other applications do need them. This cooperative memory management insures available pages will be there if needed.

My only gripe is that a color-coded memory-map diagram of all the paging areas (free page, zero page, and standby page) and RAM memory areas is not included so one can see how everything overlaps at a glance. You basically have to draw this diagram yourself to follow the article.

Post New Requests Here / Re: IDEA: TV tuner software

« Last post by superticker on November 20, 2006, 06:49 PM »

So the problem is with the programmer's implementation of ATI's MMC video recording software, not the design of Windows. Interesting. Maybe I need to look at getting new video recording software rather than changing my disk hardware configuration.
-superticker
Well, you can't always know beforehand how large the file needs to be. But you can reserve some minimum size, and then make sure you write largish buffers.
-f0dder (November 20, 2006, 05:50 PM)

Yes, but you know the recorded programs will have 30-minute quantum sizes and you can compute the required contiguous blocks from that reliably plus a little extra. Then you can release the little extra if you don't need it. That's how MVS (IBM main-frame OS) typically does it.

If you have a partition that's mostly used for large files, you can reduce fragmentation by increasing the cluster size. If you almost solely use it for video editing purposes, you might want to try going as high as 64KB cluster size.
-f0dder (November 20, 2006, 05:50 PM)

I seriously thought about doing exactly that, and if I knew fragmentation would be a serious problem, I would have done that when installing ATI's MMC recording software. But the better solution would be to find a smarter video writing application that preallocates extents properly.

Using the right cluster size is an issue. You need to fit about 5 clusters on the disk's on-disk cache and you want to have about 5-8 files open at a time, so 5x8=40 clusters needs to fit on the disk's on-disk cache. Unfortunately, a standard non-media-server disk has a smaller on-disk cache. That limits your cluster size.

Many server-oriented RAID controllers lets you specify a look-ahead buffer size for your sequential-file media disk arrays. The problem is that this setting affects every disk on that RAID system. As a result, you have to place your media server disks on an independent RAID controller from your system device, which I don't like.

But my home entertainment system doesn't use RAID. That's just too much. I want to keep my home system as simple as possible. When I replace its disk, I'm going to create a separate large-cluster-size partition as you suggested just for recorded programming.

Post New Requests Here / Re: IDEA: TV tuner software

« Last post by superticker on November 20, 2006, 10:16 AM »

Doesn't the Win32 API support disk block preallocations?
-superticker (November 19, 2006, 11:58 AM)

SetFilePointer() to the expected output file size, SetEndOfFile(), and SetFilePointer() back to offset 0. Presto, instant action on NTFS, and as few fragments as possible.
-f0dder (November 19, 2006, 04:22 PM)

Thanks. I'm assuming these calls will work for sequential access files (as well as random block access).

So the problem is with the programmer's implementation of ATI's MMC video recording software, not the design of Windows. Interesting. Maybe I need to look at getting new video recording software rather than changing my disk hardware configuration.

What bothers me most is that the 2000 extents for a one hour (1 GByte) program are scattered all over the place on the drive. The resulting slow response causes about 1% of the frames to be dropped even on moderate compression modes. One would think the video recording programmers would have preallocated each file extent in 15- or 30-minute program segments (256 MByte contiguous blocks) to reduce file fragmentation, but they didn't.

Is it fair to assume most other video recording software preallocates large extents (contiguous blocks) to reduce fragmentation and speed disk access as one would expect?

Post New Requests Here / Re: IDEA: TV tuner software

« Last post by superticker on November 19, 2006, 11:58 AM »

Being new to the TV tuner world in 2002, I was very concerned placing a TV tuner card on the PCI bus would eat up all the bus bandwidth since the tuner card would be constantly sending data to the video card. Not wanting to take a chance, I bought an ATI 8500DV which is a combined video and tuner card so the PCI bus stays out of the loop.

I've been happy with my ATI 8500DV tuner/video card over the years, although a real TV tuner may have slightly better FM sensitivity. But I always wondered if having separate tuner and video cards would really slow the PCI bus down as much as I feared? And now with the PCI Express bus, is this bus contention even that much of an issue today?

For my next video card, I've looking for a dual-tuner PCI Express bus card. Again, I favor putting the tuner and video card on the same card if possible.

---------

I definitely find the ATI MMC video recording software to be slow. The main reason for this is because it fails to preallocate file extensions appropriately. Doesn't the Win32 API support disk block preallocations? Gee, a recorded program often has 2000 file extents--that's ridiculous for sequential file access! Moreover, the cluster size on my disk is only 4KBytes.

Okay, I did think about buying a second drive that's SCSI (not parallel EIDE) and formatting it to a media-server cluster size of 2MBytes or so. Is that what everyone else does? Is a media server configuration typically used for "home entertainment" PCs? Do people also buy media-server hard drives with 128MByte on-disk caches for their large sequential access needs? What cluster sizes are you formatting these media beasts with?

I just wish the ATI MMC video recorder would better preallocate it's file extents. 2000 extents on a single sequential file is just plan ludicrous. Something is wrong here.

Living Room / Re: What's better: modern built-in motherboard sound chip or old sound card?

« Last post by superticker on November 19, 2006, 10:23 AM »

my motherboard decided to help me solve this dilemma by not working..
so now I'm off to buy a sound card..
-mouser (November 19, 2006, 08:14 AM)

You never said whether or not it's a hardware or software problem. 95% of the time it's a software problem. Open the Sound control panel and click the hardware tab. Be sure all the codec kits that should be there are there. If you're in doubt, reinstall the kits from the CD that came with your motherboard. (You can also check the registry if you know what you're looking for.)

Moreover, open up each kit and verify all its driver components are running right. Double click on driver details to see their properties. See if all the drivers (*.sys) are signed. Although they may not have been signed originally, there are probably signed versions available by now. If you think there's driver corruption and you haven't ran chkdsk recently, do it now. Any disk with more than 7% bad sectors should be discarded; it's got mechanical problems.

If it really is a hardware problem, chances are you'll get a little something out of it--maybe a crackle.

Pages: prev • 1 2 [3] 4 5 6 • next