Design and Architecture

Sounds like something that can only be done from the admin level. "The clue train passed his station without stopping." - John Simmons / outlaw programmer

Tristan Rhodes wrote: Does that sound about right? Realize this issue is about Object Oriented Design. For most of us we are students for a long time, maybe forever. Don’t expect to obtain a 100% understanding in a few days. In the concrete example you posted you have focused on the idea that exposing the data through a setter is where the design goes off track. It is an indication of a poor design, it’s not “the” problem. A more complete analysis is required. What is the purpose of the Validator class? Having a class that just determines the Boolean state and then simply exposes the state might indicate a larger design issue. Other elements of the design must be considered to know the answer.

:omg: The 60 minutes thing was a hypothetical situation to help define what I was trying to accomplish, it had nothing to do with my current solution. I wasn’t seeking critiques on my current solution only ideas for a better/different method of achieving the same goal. As always, I’m getting responses on everything except the question at hand. The question at hand being, "I’m seeking ideas for an efficient technique to sequence transactions in a SQLServer table for consumption by a serviced component app-server running on several physical machines"

You should look at: http://en.wikipedia.org/wiki/Design\_Patterns and I think this tool will help you: http://dpatoolkit.sourceforge.net Regards, Andrew

Hi Mike, Thanks, yes, that helped. I've been doing some more research into MVC and guess i originally misinterpreted it. I've now separated the model from the view and the controller is the host form. So what i have ended up with is: * A Generic Tree that raises events on change and can be serialized * A composite control set that takes an instance of the above Tree and observes it, keeping synchronized. This custom control also raises user events. * A Controller that handles all events from all controls, and manipulates the Generic Tree; cascading changes to any observing views. That seems to be a really nice way of doing things, and I'm really happy with it. I feel as though the penny has finally dropped for that particular way of thinking. Cheers :D Tris Question: Does the fact that the composite control also raises user events change its role in anyway? Or is that perfectly acceptable? ------------------------------- Carrier Bags - 21st Century Tumbleweed.

Good, nice to heard that.

may be you can have a look at Microsoft's Asyncronous block from MS patterns and practices. not sure. cheers varun.y.sharma@gmail.com

I think you should seperate the functionality from the data objects, similar to the way you would seperate database records from the actual database provider logic / queries. Instead, implement a renderer base class that takes the above primitives, then derive that for OpenGL / D3D. That way you could do: Renderer r = new D3DRenderer(); r.Render(myModels); Hope that helps Tris ------------------------------- Carrier Bags - 21st Century Tumbleweed.

Hi, the general question you ask needs an entire book to contain a general answer. Here are some facts: 1. you are in the wrong forum, it is not a software design/archiotecture issue, it depends foremost on the specific hardware. 2. Most microprocessors have some kind of NOP instruction, and can execute that in 1 cycle. So a 1GHz CPU (single core, no hyperthreading, ...) could execute 1 billion of those in one second. 3. More complex instructions may/will take longer; more so for CISC than for RISC. A divide instruction may take 35 cycles or so (or not be available at all). 4. Instruction scheduling nowadays complicates matter a lot: instructions don't always get executed at maximum theoretical speed due to scheduling constraints. 5. Memory bandwidth issues may slow things down (code cache misses, data cache misses, table walk delays) 6. Other system activity may reduce the throughput you are interested in: e.g. interrupts dealing with timer ticks, network traffic, ... :) Luc Pattyn [Forum Guidelines] [My Articles] this months tips: - use PRE tags to preserve formatting when showing multi-line code snippets - before you ask a question here, search CodeProject, then Google

Thanks Michael :) Mark Churchill Director Dunn & Churchill

led mike, Thanks for your opinion. I appreciate your comment on how much work this would take. Finding out what it would take to make such a tool is actually my main objective. You have confirmed my impression that it would take a lot of man hours. My need for such a tool is always brought to mind but never executed. I think it is going to be one of those things that will always be on the back burner. Respectfully, Mike Santos

egottwald wrote: www.winmerge.org Yes, I have, but I think the Perl utility I've just found, csvdiff[^], will get me further, and I stand to learn some Perl if I need to tweak it. MY BLOG

Well, the project is on hold but here's what I've come up with so far: You store a plaintext string (i.e. "helloworld" in the web.config) You then setup IIS so that the application doesn't recycle memory. You store a general decryption key in AppState of the application (essentially keeping the key only in memory) You use the decryption key to take your plaintext string and encrypt it (so "helloworld" becomes whatever it would be with your AES encrypted string turns it into, with or without a salt) and then you take THAT and place it into the web.config. This way, the actual key isn't stored on the server, the config files or in the database. I do know, however that microsoft has a tool for encrypting parts of your web.config, but I haven't looked into it (and only discovered it after the architecture I had just proposed.) When the application first loads, it checks to see if the key in AppState can encrypt "helloworld" to match the encrypted string in the web.config. In terms of assigning permissions to different users, you could use this same key to encrypt all the strings in the database, including a 1-many mapping of passwords to users with permission to see them. let me know if you come up with anything better!

I've actually since left the company & didn't end up using it much after all. Pity tho, as it does seem to have potential. The basic "gist" of it (for want of a better word) is that it'd allow you independence from any particular RDBMS (which is particularly useful to small ISV's I'd say). That's the upside - the downside is that it does require a significant time investment to get a decent, reusable DAL API that you could plug in to the various incarnations of one's project. The reason I didn't get to use it as much as I would have liked to is that the project I was working on was way behind schedule and my team wasn't afforded enough time to get it (an iBatis DAL) up & running. Besides that, the company decided that they would be working exclusively with SQL Server in the future, and as such, the abstraction provided by using iBatis didn't provide any significant benefits. I ended up building a lightweight DAL that could be applied to pretty much any SQL Server database with only a little extra code involved (setting up SQL parameters for specific objects). Other than that it does the job pretty well. Good article fodder were I eloquent enough :) But fortunately we have the nanny-state politicians who can step in to protect us poor stupid consumers, most of whom would not know a JVM from a frozen chicken. Bruce Pierson Because programming is an art, not a science. Marc Clifton

No, I didn't found anything yet :(

Which language are you developing with ? EDIT: if you are using Java, I suggest you take a look at OSGi[^]. The plug-in system of Eclipse is built on top of OSGi. Cédric Moonen Software developer Charting control [v1.2]

Google around. I had a link somewhere, I wish I could find it :doh: "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon

Software Quality Metrics is pretty broad topic and finding anything free will be tough. "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon

Nope. Don't have a need for it. Thanks. "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon