ATL / WTL / STL

19-May-13 15:55

I am using Visual Studio 2008 on a Windows 7 machine to generate a utility to validate TMATS and Chapter 10 file. These are standardized files used in telemetry systems. The files begin with a text based description of the data (7 bit ASCII) followed by large amounts of telemetry data in binary format. The binary data will not be shown, but there are some validation checks to be done.

In the text based preamble there are about ten different major sections, some rather long. Each section describes some aspect of the data. The basic format is an identifier that ends with a colon, followed by text that ends with a semicolon. The next identifier follows.

My plan is a separate dialog or window for each. The user will be able to edit the data and write out to a new file.

I am unsure as to the best solution type. So far I am leaning towards an MFC application with single document support. (Only one file will be open for reading only, and a new file optionally opened for writing.) The next best option seems to be with Dialog support. Or maybe someone will suggest something else completely.

This will be open source code posted on my web site for others in the telemetry industry to use and comment on. I plan to make all the classes and code as simple as I can so it can be ported to other systems. But I will not go to extreme lengths for portability.

I am fairly good at writing code, but on the lower end of intermediate for Visual Studio.

Pleas make your suggestions.

Thanks for your time
If you work with telemetry, please check this bulletin board: http://www.bkelly.ws/irig_106/

Garth J Lancaster19-May-13 17:49

19-May-13 17:49

bkelly13 wrote:
classes and code as simple as I can so it can be ported to other systems

hmm, I think that rules out MFC ..

Seriously .. I would first abstract out what you need to do with a file, to get your sections etc - this might be one class or dll ..then I'd look at the UI - why dont you grab a copy of balsamiq for example and mock up possible UI's - or even resort to plain pencil and paper sketches

Once you've got the abstraction/logical stuff on paper, then worry about what you're going to build it in

'g'

bkelly1320-May-13 15:17

20-May-13 15:17

Hello Garth,
Yes, I have a pretty darn good idea of how I want the data displayed. I have done a lot of thinking on this project. If you think the desire for portability rules out MFC, then what does it rule in?

Thanks for your time
If you work with telemetry, please check this bulletin board: http://www.bkelly.ws/irig_106/

Garth J Lancaster20-May-13 15:26

20-May-13 15:26

well, depends on 'how much portability' - I thought of 'Qt' and C#/Mono first off .. if you go the C#/Mono route I think that gives you winforms on linux (Mono)

'g'

bkelly1320-May-13 15:54

20-May-13 15:54

I have visual studio and home and at work, but neither of the suggestions you made. What would you recommend within VS?

Thanks for your time
If you work with telemetry, please check this bulletin board: http://www.bkelly.ws/irig_106/

Garth J Lancaster20-May-13 16:01

20-May-13 16:01

not sure about the 'within vs' .. but this might help http://www.mono-project.com/Guidelines%3aApplication_Portability[^]

bkelly1320-May-13 16:12

20-May-13 16:12

VS for Visual Studio

We have only Windows where I work. I do not expect the code to simply compile and run on other operating systems. I just want to ease the task a bit for someone that might want to do that.

So which Visual Studio Template do you suggest for this project?

Thanks for your time
If you work with telemetry, please check this bulletin board: http://www.bkelly.ws/irig_106/

Garth J Lancaster20-May-13 16:20

20-May-13 16:20

nooo .. you dont say

I think if you stick with a pure VS/Microsoft environment, your portability options are limited, again, depending on 'what portability' you want .. what I was thinking of with the Mono stuff is, and I'd need to check, it allows you to have pretty well one code-base - you'd just have to compile it for mono on the Linux platform

Why not write it in VS/MFC now, the core functionality (ie, not the display stuff) can be separated out into its own layer, and then consider where else/what else (if ??)you need it to run on...

There are 'Open IDE's for c, c++ iirc, that work with VS compiler, GCC etc (although the name of one eludes me right at this second)

bkelly1320-May-13 16:27

20-May-13 16:27

Re: nooo ... you dont say.

Re: Re: not sure about the 'within vs'

That sounded to me like you did not know what I intended with "vs." I did not mean to be sarcastic or anything.

Within MFC: I read about the various options: multi document, single document, and dialog in the help file, but it was really not very informative. Single document and multi document both provide a window with File open menus and the such already added. Dialog does not have that, but it "appears to be" a much simpler project. I only need to have one document open at a time.

Keeping this simple: There are about ten sections to a TMATS header. To make it simple for the reader I want to put the data for each major section in a separate dialog/window/(or some silly entity).

So which MFC option would you take?

Thanks for your time
If you work with telemetry, please check this bulletin board: http://www.bkelly.ws/irig_106/

Garth J Lancaster20-May-13 16:32

20-May-13 16:32

I would probably go Dialog, or Single document,

thought 1 : a tab bar, each tab being a section/TMATS header

thought 2 : 2 'panes' - left hand 'pane' is a tree structure (or listbox ?) of the file sections, when clicked upon, the TMATS header is displayed in detail on the right hand 'pane'

bkelly1320-May-13 16:38

20-May-13 16:38

This is really helpful for me. I would not have thought of either one. If I google each will I find a good description?

I think the multiple tabs would be easier, but the split pane concept (kind of like Windows Explorer?) would also be cool. What is the approximate work effort for each? Maybe on a one to ten scale? I don't really care what the ten on that scale really means, just a comparison between the two. Maybe assign ten to the more difficult one and assign something less than ten to the easier one.

Thanks for your time
If you work with telemetry, please check this bulletin board: http://www.bkelly.ws/irig_106/

Garth J Lancaster20-May-13 16:49

20-May-13 16:49

just on a quick thought, the tabs are easier - caveat : as long as you only use them for 'displaying' data, ie, you dont have things going on in tabs that you need to pass between/to other tabs (Ive used a 5 tab bar as a workflow type thing, but an error in tab 3 meant they had to go back to tab 2 for example). One thing that may go against this is the size of the dialog, vs the number of tabs - 'fitting them on/scrolling them' so all the info is apparent to the user - with the dual pane view, you might not have that issue.

Im thinking '4' for tabs, and maybe '7' for split-pane

Again, if you have a class, with unit tests etc, that can handle parsing, 'getting sections', getting data from your TMATS files, then fitting it with a new UI shouldnt be a biggie

Im pretty sure google would bring up plenty of articles .. there's bound to be some lurking here on cp, if only for ideas

bkelly1320-May-13 16:59

20-May-13 16:59

Thinking of Tabs: If I have a class in charge of writing, can it access the tabs to fetch the data, then write the data to a new file?

I presume I would need to create some type of writer class that would be able to poll each tab asking for the data contained within the tab. (Presuming each tab is a separate object ) to provide the data to the file writer. Then the file writer class would poll each tab to get the data to write out.

Its past my bedtime so I'll not respond again and catch up tomorrow. But again, this has been very helpful for me. Thank you for taking the time to maintain this conversation.

Thanks for your time
If you work with telemetry, please check this bulletin board: http://www.bkelly.ws/irig_106/

Garth J Lancaster21-May-13 2:16