Disney + Marvel

I wonder what this will do for the Amalgam universe?

http://www.marketwatch.com/story/disney-to-acquire-marvel-entertainment-2009-08-31-9050

Precision

I used to argue with my dad on the matter of precision. My dad is and engineer in education and personality as was his dad and his dad's dad and so on, so I inherited the mentality critical thinking and interest in mechanics, processes and technology of an that such a personality brings (Which sadly doesn't help with personal relationships:).

At one point I had decided I wanted to design my own game engine. I had decided I wanted to use the most precise decimal system I could find in 'cough' VB for storing the locations of points. In the process of designing the basics of the engine I tried to find pi to the most decimals that would fit in the desired unit. I had learned the 22/7 trick from my Drafting instructor but was completely unsatisfied with the accuracy of only 2 decimal points.

My dad asked why I thought I needed any precision beyond two decimal points; in the mechanical engineering world two or three decimals are usually considered accurate enough for almost any consumer products, as anything more precise makes little difference, and just the expansion and contraction of materials alone exceeds the difference such precision would bring. 

But I was using a computer and despite what the 1950s and the movie industry hope to tell you, computers are inherently inaccurate at least when it comes to trying to calculate decimal and floating point numbers.

If you use floats for storing points in a world you can move around in, you would find that the precision of the locations of points in the center of the world is quite high, but the farther away you move from the center the less precision you have because the whole number portion of the value is getting filled with the larger numbers sacrificing the precision and storage space of the decimal portion. So at the edge of a large map you could end up with unexpected gaps in your world.

Well, okay if you were to only use integers the gaps could be reduced, but when you want to try to copy the nature of your surroundings you might find you want more precision. The thought might come up to use fixed point, sure it could work but you would have to be very careful with your rounding and your geometry.  You also end up with geometry that can look a little blocky depending on the resolution of your values.

But I never got far enough along to be able to see any degradation. So now days, I've stopped caring. Well almost, when I go back and play with old geometry code I am still tempted to write everything as doubles rather than singles, but now I've learned to think processor capacity and the effects on speed too.

Batch Script Search and Replace

A while back I thought I would try to share some of my favorite batch script snippets from my Google notebook but the shared notebook was quickly flagged as potentially dangerous and I didn't feel like contesting the flag so I pulled the snippets back into my private notebook.

I still want to share a few of my most useful snippets, so here is my favorite which I have heavily adapted it to fit many needs

Replace.bat

@Echo Off

Set /p Source=Text you would like to search:
Set /p Search=Text you would like to find:
Set /p Replace=Text to replace "%Search%" with:
Set Dest=

Call :Replacer "%Source%" "%Search%" "%Replace%" "Dest"

@Echo.%Dest%

GoTo :EOF

:Replacer
    Set t=%~1
    Call Set t=%%t:%~2=%~3%%
    Set %~4=%t%
GoTo :EOF

So what's going on here?

Well, first you provide a string to search, what you are looking for, then what you want to replace it with.  The script then calls a function which replaces all instances of the search term with the desired replace term. Simple right?

The interesting part comes in the line which does the search and replace.

variable t is a temporary variable which gets set as the source string then it is replaced with the new string in a call set statement in this strange part "Call Set t=%%t:%~2=%~3%%" which could be written as "Call Set Destination=%%t:%Search%=%Replace%%%".

What it expands to is fascinating as it double expands. Say you want to replace "is" with "was" in the string "This is your search string". You would end up with "Call Set t=%%This is your search string:is=was%%" which needs to be expanded one more time as it is now an environment variable replacer which the call statement executes to replace the instance of "is" with "was" so you end up with "This was your search string".

The final confusing piece of code sets the result to an arbitrary variable name which you feed the function in this case "Dest", this allows you to reuse the code with any destination variable you would like.

This has been a fascinating  howbeit confusing piece of batch script that I have found to be very useful in so many situations.

Now for a few notes about batch code hygiene in which this sample employs.

1. Never wrap strings used in set statements with quotes "". doing so will make string cleanup difficult. Set will set a variable to the entire line after the equals = sign.
2. In batch, all variables are global, so work with them accordingly. You may end up with confusing results if you don't understand this.
3. Create empty variables for readability. If you have a set of variables where one will be set later but you can set it blank near the top then fill it later, this is only for readability.
4. When you pass a string to a function, even if you are passing a variable, quote "" it. You never know when you will be passing a string with spaces or some other weird characters, so in order to make sure it is all kept together quote it.
5. When you retrieve a variable from a function use the tilde ~ eg. %~1. the tilde removes the surrounding quotes from a string passed through a function.
6. When echoing a variable use "@Echo.". This prevents displaying the echo state such as "Echo off" if the variable ends up being empty. The @ prevents echoing the command if Eco is on.

Photo Tags

There are two graphics applications I use on occasion which automatically scan through photos and identify faces. They are Google's Picasa and Microsoft's Windows Live Photo Gallery. They both have their own unique uses and they are both getting better, but every time I move to a different computer or install a new operating system I find myself doing the same thing all over again.

On my little network hard drive I have tens of thousands of photos must of which are of people. Right now I can remember who many of them are, some I have forgotten their names already, but I want to make sure I keep track of who the people are. I've started a few times tagging people and adding comments to pictures, but then I go on to another computer only to discover the information was stored on the first machine only.

Windows Live Photo Gallery has built in tagging of people which is nice, the auto identification could be improved however. The manual tagging is almost as good as FaceBook. Actually it would be nice to be able to sync to FaceBook and other accounts to help identify who someone is, as it is now the only option is a Windows Live account.

Picasa doesn't seem to be able to locate people using just the installed application, but it can when the photo is put online and the identification online is actually pretty good, but I don't want to put all of my pictures online, I don't have enough space available for that on my account for that and I'm not wiling to shell out to get more right now.

What I wish would happen is they would write the tags directly into the photo's metadata so that it travels with the photo. And it would be nice if they used a method that would be interoperable between applications so that you could identify someone in one application and it could be associated as the same person in another application. Okay, that might be a near impossible idea but it would sure be nice.

Make Office Better

Okay, I use Office quite a bit. At work Outlook, Excel and IE are my main tools, at home Word and Excel are among the top applications in my list. I write stories, keep track of my finances and hours and so on in these applications.

So when I found out that a couple of guys on the Office team decided to create a web page asking how Office could be improved; I jumped at the chance to try to get some of those pesky irritations I keep running into fixed in some future version.

So if you want to make Office better, just head on down to http://makeofficebetter.com to send in ideas suggestions and vote on things you want to see fixed.

OS's and their Shells

I've been thinking (which is probably a dangerous thing) about what an operating system could look like if you could just start from scratch.

What if you could just ignore the existence of the current interface paradigms and created something new? What could it look like?

The current windowed paradigm goes all the way back to the old Xerox prototypes and everyone has been copying and improving on the idea ever since Apple, Microsoft, Unix, Linux, and the list goes on. Just about all mainline operating systems today package the jumbles of compiled code in boxes and call them programs.

I suppose unique is what you have with Singularity; I don't think I have seen an operating system that looks like it before, though it is pretty much all command line with a Matrix like feel. It's cool but most of the coolness is hidden behind the shell.

A cool design I have played with a little around work it the Surface PC which lends it's self immediately to intuitive interface design.

So, if you could design an OS and it's shell, what would it look like? Would it be 2D, 3D, Command line… what?

Watching what is going on in HTML development it might not be long until we have 3D web. Could it be the next OS interface design?

Would you have a game shell with lots of fancy graphics and cool effects or would you have a "keep it simple silly" style to boost processing to things that don't render on the machine to make a work server?

Hollywood has tried time and time again but the shells that are designed for TV and Movies are ridiculous at best designed for flashy effects rather than actual use. But they do present new ideas from time to time.

I'm not sure what I would do. The possibilities are limitless. I think 3D should be a native API, but not sure how it should be used.

Anyway, food for thought.

One Tuesday Down

Alright, in three weeks I can start applying for jobs at Microsoft again which will give me five weeks that I will likely be interviewing about three times a week until the hundred days is up on the 8th of October. That means that I have three weeks of nothing to fill.

I did sign up for a two day ASP.Net class next week hopefully it is interesting.

Today After watching a little of Season 4 of Alf I couldn't stand being cooped up anymore and finally got out of the house and rode my bike around Robinswood park for about half an hour then down to the Chevron near Albertsons to fill up my tires (It's the only place I have found around here with free air).

After that institute. The lesson was continuing the previous subject and adding priorities and goals.

Classes vs. Structs

I'm having a hard time getting to sleep tonight (probably too much sugar in the strawberry milk) so I thought I would write out what is on my mind.

Classes vs. Structs.

In C# and VB.Net there are both Classes and Structures which can act fairly similarly, but the have differences that make one more appropriate in a situation and the other in another situation. For my little geometry program I have been debating back and forth on which to use for the shapes.

So far I have used Structs as it just seems natural that they should be. The other advantages is the natural garbage collection, and that they are initiated on the stack rather than the heap so they are lighter to fling around.

But there are disadvantages for example not being able to use inherited classes. This is what I'm debating over.

Classes are a little heavier and tend to be initiated on the heap so there is possibility that there could be some lag to their use. But they can Inherit other classes which means that you can create a base class with all of the display stuff packaged up into the class which other classes can inherit and use like it is their own. The other classes could take care of all of the other weird stuff but retain the same basic variables.

On the other hand both classes and structs can use interfaces. Interfaces are similar to inheritable classes except that they only define what a class or struct must contain to be interfaced as it's type.  This means you have one interface with a whole bunch of heads defined but nothing else. It also means that in order for a class or struct to be defined as that type it must contain full definitions for those pieces which the interface defines. This can bloat the code a little.

So, Define them like I already have as structs and continue adding the rendering and graphical property code, or redefine as classes and split off the graphical property code to an inheritable class while retaining the rendering code in the interface?

I've only been thinking it through so far, the idea seems plausible classes might offer the advanced flexibility I'm looking for. But I'll have to see if such a hybrid works when I start playing with the code again.

RIP MS Money

Soon after I moved up here to Washington I was given MS Money as a gift to help me maintain my finances. It does budgeting, spending tracking, home inventory, bill tracking and so on. But for the most part I use it so see when Paychecks and debit transactions have been processed and to see the current "actual" balance of my accounts. Basically I'm using it as a check register.

I then copy the transactions into an Excel Spreadsheet so I can see the pretty line graph which helps me understand how much I actually have and what is available since the numbers mean little to me.

Well… My 2 year online update license ends next month and Microsoft has abandoned the entire line stating that as of June 30, 2009 it is no longer available for purchase and those who were able to get the updated version before that date would only be able to use the online resources until January 31, 2011 when they will be taken down. Though I did notice a new note dated August 5, 2009 that there will be a version of Money that will be released that won't require activation, but also won't have online features.

So I'm left trying to figure out what to do. Microsoft is working with the makers of Quicken to improve importing as an option. But I'm not so sure I want to use Quicken.

I have data that goes back to August 16, 2005 (I know, not all that far back but still) data which is no longer available from my financial institutions so what ever I use I want to keep that data.

What are my options?

• Intuit's Quicken. But do I really want a new desktop app?
• Quicken Online. It is supposed to be free using advertizing. Not sure more about what it is though.
• GNUCash, just doesn't look like it is up to handling it all.
• mint.com. sounds like it could work though I'm a little worried about giving them my account information. I can't import old data that is no longer available. Hmm. But it would alert me when the new transactions have completed. Not having the option to download data seems sad.
• justthrive.com. It's supposed to be similar to mint.com.
• wesabe.com They don't require account information so you have to enter it all manually. Sounds like a lot of work. I'm a programmer and therefore am lazy by nature (Little joke) I am the type who will do extra work to automate and simplify to reduce later work.
• clearcheckbook.com. Looks similar to wesabe including ading and importing.

I'm not so sure I want to trial a whole lot of these to find one or a combination that works right.

Then again, what if I just wrote up my own financial software and used some free service to monitor? Eh. Time will tell.

Game of Life

Conway's game of life is an amazingly addictive game http://www.ibiblio.org/lifepatterns/ or http://conwaylife.com/ the concept seems so simple but it creates such fascinating complexity. It's especially fun finding the patterns that actually increase indefinitely.

Browsers

Watching the web browser statistics is fascinating. In one glance you can get an idea of what people are using to view the web and get an idea of what browsers to focus on for testing for the majority case. It is also interesting to see how a particular browser is being accepted by the community. The one I have been watching lately for the general case is StatCounter(http://gs.statcounter.com/#browser_version-US-daily-20080701-20090731) which shows a nice graph that I can watch.

It is fascinating to see that IE6 is still hanging in there. IE7 held high ground for quite awhile there but it looks like it is slowly being replaced by IE8 now. I'm a little surprised to see that FireFox is starting to loose ground. And according to the chart Chrome is out of the picture so I'm guessing the latest version of Chrome is actually part of the confusingly high Other category.

Now that was by version number, which does not show the brand loyalty. Looking at StatCounter by Browser brand (http://gs.statcounter.com/#browser-US-daily-20080701-20090726) IE remains in the lead but Firefox pretty much mirrors IE across the 50% line which is a fascinating trend.

The rest look like they are staying pretty much even except for Chrome which is slowly gaining ground, however it is hard to tell but it almost looks like Chrome is getting customers from FireFox.

Anyway fascinating statistics. It will be interesting to see what happens when IE9 comes out in a couple of years.

Strange Hobby

Every once in a while I will get frustrated with a website that I try to use and I send feedback in the form of a Bug report:

Descriptive Bug Title

------------------------------------

Short description of problem.

1. Repro Steps:
2. First thing to do
3. Second thing to do
4. And so on…

Expected:

The expected behavior.

Actual:

What actually happens when the repro steps are followed.

Recommendation:

Any recommendations to correct the issue if available.

Remembering the Domain Name Boom

I remember back to when the Internet was still new and domain names were being registered for the first time it was fascinating to watch as word after word was being gobbled up to assign new pages to. It looked like it wouldn't take all that long before every word in the dictionary and a huge portion of short phrases would be claimed. I figured that by the time I was an Adult they would all be claimed.

Not having internet connection at that time I used to write down every web address I wanted to visit on little sticky notes which I kept inside the flap of a calculator that I used to carry around, in the hopes that I might be able to spend a little time at a library, lab or somewhere else looking at these fascinating pages. Oh how things have changed.

So it is interesting then to see the results as an Adult now when indeed a good portion of the simple addresses have been claimed but there  are still words and phrases left untouched and many of the pages once claimed have been left untended abandoned or taken down. Some domains have even been let to expire.

I wonder whet the next decade will bring?

Windows 7 Has Gone Gold

Woo hoo! Windows 7 was signed off to RTM today!

He’s Walking a Tortoise‽

Had to make a double take on the way home from work today. At one of the corners near where I live there was a guy walking a Tortoise down the sidewalk.

Management

I’ll just say this about the past week at work: I sure don’t envy Managers jobs. Talk about stress.

Replacement

Well, tomorrow I start training my replacement. I'll have a month to start winding down the work load and start transferring my tasks to her. So It's that time to make sure everything is well documented so someone new can try to pick up from where I left off. It's time to make sure the few tools I've created work perfectly so that they just work for whoever uses them next. Soon it will be time to say goodby for another 100 days. Now to figure out what to do for the 3 months I'll be without work.

Remote Control

Woo hoo! I’ve finally got Windows 7 RC installed on both of my machines at home and I have been able to play around with Homegroup a little. It’s kind of cool. Since I have been using Win 7  at home, when I go to work and back to Server 2008 and Vista I have to remember that I can’t do some of the new cool things like maximize a window only vertically or drag to un-maximize.

One of my machines is hooked up to the TV so that I can watch a few sows I like on hulu. But it started getting frustrating getting up and down to switch to other shows, or whatever.

So I started looking into ways to control that machine. I tried Remote desktop since it was already on the machine, but it would log the console account out when I logged in, so it can’t work.

Next I tried Remote Assistance since it was also already installed but it took going back and forth between between machines a couple of times to gain control, and the shortcuts were only good for a short amount of time. The cool part was that I could actually see both screens on that computer in the remote assistance window, and since the the TV is rather fuzzy at computer resolutions, it was nice to be able to read the screen in the window. It is nice but limited.

But being able to see the screen locally isn’t necessary. What if I could just move the cursor to the other machine? I heard of a program at work which did just that but I hadn’t caught the name of the program. So at home I did a search and found Synergy, which is just the program I had been hearing of, a virtual KM switch. So far it is working nicely, I’ll have to see how it works over the long term.

Anyone?

On Monday my manager’s manager came into the office saying he had been approved for contract overlap and asked who they would get to replace me. I said I don’t know. He then said he was serious and asked if I knew anyone who could take over then told me to think about it. Hmm. I still don’t know.

[Update]: Well, it looks like my replacement has been hired. I meet her on Monday morning and so that leaves me a month to train her on everything she needs to know to do my job.

HTML Confusion

Hmm. Interesting. The following HTML does not behave as I would expect it to in all of the major browsers:

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html>
    <head>
        <title>This is a test</title>
    </head>
    <body>
        <div>
            <a href=""/>This should not be a link.  
            <b />This should not be Bold.  
            <i />This should not be Italic. 
            <u />This should not be Underlined.
            <div style="background: red;" />This should not be Red. 
        <div>
    </body>
</html>

In xml, a tag with the form <x /> would be an empty xml element. So the rendering of the above ends up being quite confusing as each of those tags ends up being opening tags without closing tags.

Perhaps this is some remnant of the pre-xml days when html was still being defined.