Entries for John Melville-- MD

Can I launch a program when a FOLDER is double clicked [Can I launch a program when a FOLDER is double clicked]

John Melville-- MD — Thu, 01 Oct 2009 03:19:15 GMT

I write software to assist medical professionals in collecting evidence in cases where children may have been abused. I have an interesting problem with data storage.

I would like to have my "data file" actually be a folder. My files are large (200+MB) and contain multiple JPEG, WMA, and (soon) WMV streams. I would like to use the file system to manage this stucture because that's what the file system does best.

Furthermore, when my district attorney discloses my data file to the defense (like we always have to) I want it to show up as a completely normal folder so that the defense experts (who presumably don't use my software yet) can quickly see that its just a directory structure full of file formats they already know about. Since we always disclose evidence on read only media I am not worried about the defense modifying the folder.

I am worried, however, that my own users will muck with the files in the folder, and my software is not robust to this possibility. (Right now I am using a SQL CE database that is pretty hard to modfy outside of my software.) I also want to have a real easy "just click on the file and my software opens up" experience.

Is there any way to make a folder look like a file? By this I mean that when you click the folder in the shell my software would open up instead of the shell opening a new shell window with the contents of the folder. It would be very nice if I could replace the folder icon with my own, this would make the disguise sufficent that any users who manage to open the folder in the shell are sufficently skilled to be held responsible for their own misery.

Thanks in advance for any assistance

in reply to Can I launch a program when a FOLDER is double clicked

Bitmapeffects reprocessed sample [Bitmapeffects reprocessed sample]

John Melville-- MD — Tue, 16 Jan 2007 02:18:12 GMT

I need to write a custom Bitmapeffect for a project I am working on and have sound a surprising absence of samples and details about how to make anything work. Even more depressing is that even after I found RGBFilterEffect in the SDK, it didn't compile, and then it crashed, and then it required com registration, and on and on.

Lou Franco presents an interesting idea that branches off the sample, write the BitmapEffect in C++/CLI and then you can package the managed and unmanaged code in a single assembly. Furthermore since the unmanaged half is living in the same assembly, a little bit of COM goo (that I frankly don't understand) removes the requirement of COM registration, and restores our favorite CLR x-copy deployment paradigm.

Unfortunately, Lou, doesn't give a complete sample and his code doesn't quite compile. I spent the better part of two days mucking around in C++/CLI, a language I don't really understand, to get everything pasted together and working. I thought I would share the result.

For those of you, like me who couldn't care less about the pipes and just want to muck with the pixels look in in the RGBFilterEffect.CPP CRGBFilterEffect::GetOutput

in reply to Bitmapeffects reprocessed sample

A new solution to Anti-Trust Version creep [A new solution to Anti-Trust Version creep]

John Melville-- MD — Mon, 16 Oct 2006 08:50:13 GMT

The EU has demanded that Windows be availaible without Media Player -- Hence the "N" versions.

South Korea has demanded a "K" version that lacks windows messanger.

The obvious version creep and testing matrix is simply getting out of control.

To effectively solve this problem I propose one cut down version to satisfy them all! Behold Windows "L" (for lame) edittion.

Its simple. Set up a website where government (or any 2 bit company) can complain about anything in windows. Say the word and Bam! its gone. Windows "L" just got a little bit lamer.

Just think about how much you could save in media and support by installing and running Windows L. Windows Lame edition: No Patchguard or Security Center for Symantec, No IE search bar for Google and Yahoo, well really no IE at all to satisfy Firefox and OS. Office Lame edition could have no PDF or XPS support, by default or for download.

We really should extend the new spirit of cooperation to all the other losers who have wasted our time in the past whining about progress. We could really make Netscape the default browser, presuming we could still find a codebase that actually compiles. Of course we couldn't actually run Netscape because there wouldn't be a GUI; a necessary concession to Apple's former legal response to Microsoft competition. I'd even remove the file system because the excellent minds at Samba couldn't reverse engineer it correctly. I hope you understand that I am really going out of my way to please everyone here.

There you have it, every legislator's dream: a completely unobjectionable version of windows and office. I think I will assist this effort by providing the entire, compilable source code, absolutely copyright and royalty free.

#include <stdio.h>

void main () {
    // printf("Hello World!");
    printf("Windows Vista L Edition Version 6.0.0.1");
    printf("C:\");
}

in reply to A new solution to Anti-Trust Version creep

FOSS and the Last Mile Problem [FOSS and the Last Mile Problem]

John Melville-- MD — Mon, 20 Jun 2005 23:31:35 GMT

Retailers, I am told, speak about the problem of "the last mile:" how do you get product from your lowest level faculity to the customer's living room? You can build a store and make your customers carry product home, you can hire UPS to deliver it for you, and there may be other options. Needless to say the "last mile" is a pesky problem on the end of a business plan and outside many retailer's expertise; but when done wrong it can condem the entire enterprise.

I contend that FOSS, speaking in general, has a problem equivilent to the final mile. The installation / first use experience for many of the OS products I have used just stinks. Half the time I have to chase down two or three dependencies from halfway arround the net. I often have to read manuals of varying quality and work through clunky user interfaves. It usually takes me an hour or two just to get the simplest FOSS program working, and decide if it meets my needs.

(At this point Beer28 will post and indicate that he just installed [insert FOSS program here] in under 15 minutes, that it worked on the first try and that the interface is better than microsoft's. Since I do not know beer, I cannot contest his experience; he likewise cannot contest mine.)

I find this interesting because I am currently working on Nester, a FOSS coverage and mutation tester for C#. Nester is right at the "one mile out point," it works great on my machine, but has bunches of dependencies, and making it work anywhere else is a challenge. Furthermore, my enthusiasm is waning. The technical problems have been solved, and the product is maximally useful to me right now. I see nothing but boring work between now and release. In a system that depends on volunteer programmers, this is a big problem.

Can FOSS overcome this phenomenon and give me the great install-and-it-works experience that I get from microsoft? Can a software system that depends on support calls for payment afford to give me that experience?

What do you think?

in reply to FOSS and the Last Mile Problem

Put Windows on a Diet [Put Windows on a Diet]

John Melville-- MD — Sat, 26 Mar 2005 21:24:51 GMT

I was scrounging for disk space and found out that my windows directory tips the scales a little heavy at 40 gb. I turned off system restore (which said it would delete all the restore points) without an appriciable gain in disk space.

I know I installed windows of a single 650 mb CD so either there is a lot of fluff in there or microsoft has found some absolutely fabulous file compression!

Anyone know how to make windows shed the extra pounds?

in reply to Put Windows on a Diet

C# 2.0 Grammar [C# 2.0 Grammar]

John Melville-- MD — Wed, 29 Sep 2004 01:53:26 GMT

Does anyone know where a grammar for C# 2.0 can be found? The 1.7 specs had a complete grammar for that version. The 2.0 specs discuss the grammar at length, but don't collect the changes in one place.

I would really appriciate any pointers on this one.

in reply to C# 2.0 Grammar