Now, it is no secret that Wi-Fi is not free at all hotspots. The most bothersome example for me, personally, is Starbucks. The closest coffee shop to my house at home is a Starbucks, so it’s a pain in the butt when I want to go somewhere to work because it means I’ve got to go across town to get to a coffee shop with free wireless internet access. From what I understand, they have started allowing individuals to comment when they purchase a “Starbucks Card”, but it’s still another hoop to jump through to get my internet access.

I am not oblivious to the fact that internet access costs businesses money. In fact, knowing the broadband access solutions that are available in most localities, I’m willing to bet that there are significant price increases over what individuals like myself pay for internet access to our homes. Additionally, business have to worry about the problem of free riders who do not pay for the product, but use the internet access. This complicates the problem when there are limited access resources available, and paying customers are not able to get onto the internet. However, if they charge money to their customers (on top of the product) odds are that a lot of people like me will opt to travel a little further rather than pay more than the cost of my drink for internet access. So, then, what does the well-meaning business do to preserve ease-of-access to customers?

Now, as I’ve implied, the win-win situation is when the customer is able to get free internet access only after purchasing some of the business’s product. My proposal is to remove the barrier between the cash register and the Wi-Fi Access Point. Consider the following:

In this scenario, each user receives a unique access code that is linked to their receipts. This means that individuals who are coming in to buy the product can access the internet while limiting the access to individuals who are free riding. There are two weaknesses that one might pose: the first is the difficulty of the actual implementation and the second is the loss of community as a result of requiring users to verify that they are customers. I will answer the question of community first.

It isn’t rocket science to figure that if a group of high school or college students are going to gather to study (cough*facebook*cough) together, it is unlikely that every individual is going to go up to the counter and buy something from the store. This is why I suggest allowing each access code to be associated with more than one device at once. So, if Fred and Sally want to study facebook together and Sally doesn’t want a drink, Sally can use Fred’s access code to gain internet access.

Now, the final point for me to discuss is the technical implementation. The University of Georgia has implemented a campus-wide wireless network system called PAWS that connects to the University’s central login system to authenticate students and grant them access to the wireless networks on campus. Without going into too much technical detail, I would suggest that the time it would take to set up a smaller scale version of this would pay for itself by encouraging individuals to be patrons at those businesses that provide free internet access.

The program is called jMemorize. If you haven’t heard of it, I highly reccomend downloading a copy and checking it out. The idea is that you create flash cards (and you can even categorize them) and then work your way through them. As you do so, the cards will move up to the higher decks, and they will take longer to “expire.” The idea is that the farther to the right that a card gets, the better you know it. This method of card sorting allows me to focus on my weakest links, if you will, when I’m getting close to crunch time. It’s really nice.

There are a few things that I’d like to see in the next version of this program though….

• Better integration with OS X. The menu items should be relocated up to the menu bar and the name of the program up at the top currently reads “jmemorize.Core.Main” instead of jMemorize.
• Auto-save after learning. All too often I will go through a lot of learning and forget to hit the save button. As a result, my progress is lost. I don’t even get a warning when I try to close the program without saving!
• Better visual integration with the host OS. The buttons and everything don’t change from platform to platform.

I’ve tried to get the jMemorize codebase up and running on my copy of NetBeans before, with no success. Hopefully, when I finally find some time on my hands – I might be able to implement some of these features myself. That is the point of Open Source isn’t it? Regardless of these minor shortcomings, I highly recommend jMemorize to anyone who needs something to help with memorization. It has a number of advanced features that make it useful for Foreign Languages, Hard Sciences, Mathematics, and virtually anything else you could make a flash card for. So, check it out!

I have successfully started transferring files of variable length from the sender to the receiver. The entire contents of the file travel successfully and correctly appear in the destination file – however, while the sender successfully realizes that the file has been transferred successfully, the receiver seems to be a little bit clueless and continues waiting for more packets. So I’ve got to track down that bug. I would like to change the protocol to include in the greeting message how many packets the receiver should expect, but I don’t think it would work with Dr. Dan’s classes if I did that. So I’ll have to figure out another way around that.

Regardless, I feel I’m making good progress on the project!

In other news I’ve started working on the second 4760 Project. The good thing is that I’ve got both sides of the program talking to each other, although not transmitting files. The receiver program is in the habit of always answering a transfer request with a NAK packet… but I think it’s good progress so far! All of my prototype code is written, so with luck testing will go by pretty fast.

The good news is that I have my project works at 100% on my local machine. Of course, as soon as I throw it on Zion, things start going wrong. I have had difficulty accurately retrieving the correct exit values from the system exec calls. I’m getting some weird results too. It either always returns an exit code of 0 (meaning that the program was executing successfully) but without giving any output, or always returning 255 (meaning that the program couldn’t run – and probably wasn’t found in the path). It’s really odd…

I think I might be taking advantage of Office Hours on Monday to get Dr. Dan to take a look at my code. I’m thinking a second set of eyes might help see something I don’t. Since I tried deploying to Zion, I’ve been attempting to develop on Zion, without the use of my usual development and debugging tools. In the worst case scenario, I’m going to have to use gdb and figure out why it’s not working… but let’s hope strategically placing couts will make due…

I’ve mainly avoided it because from what I understand, how different systems handle shared memory between multiple threads differs based on implementation. I figured since I was doing my original development on Mac OS X and then deploying my production copy to Zion it would be wise to keep it simple, even though they are both Unix derivatives when you get down to the bare metal.

My to do list at this point is to get the server executing files using the XBitHack (but Dr. Dan says he needs to tweak that part of the assignment, so I’m holding off on implementation) and then to get the server running as a daemon process. After that I should be done. Hip hip hooray!

However, I can report that I have made some progress working on the first project. I have the server listening on a socket, accepting a connection, and having an exchange with a client. Now, the project is to write a basic HTTP server, which I could probably accomplish fairly easy using straight C, but to be honest I’ve been so immersed in Object Oriented Programming that I truly think in those terms – so I’ve decided to implement using C++.

The plan is to have the main function manage the connection, then take incoming data and parse them into HTTPRequest objects then use the HTTPRequest objects to generate HTTPResponse objects that can be serialized to strings and sent back over the connection. So, while I haven’t implemented the objects yet, I have defined all of their headers and class members – so I’ve already got a pretty detailed specification of what each method will do.

I’m optimistic that once I get all the socket management stuff down in the main function then the rest of the code will be a cakewalk, because it will almost entirely be string manipulation from that point forward. Gotta love OOP.

The primary purpose of this post is to initialize the Networking category so that something exists. Expect some more exciting content later!