The problem: you want to copy some files, but only the newest and of a certain type, such as, only .txt files.

The solution: The trick is to use ls -lrt to sort by and show the date and count the number of files. Then you just pass that number to tail, and it only grabs those files! Furthermore, you can add a grep at the end to only select the files types you want.


ls -rt directory/*.txt' | tail -n `ls -lrt directory/*.txt' | grep "Nov 30" | wc -l`


Note: If you grep for a date, those files have to be the most recent in the list for this to work.

Nkurz on Hacker News Summarized it pretty well:

She teaches a class at Stanford and offers each team $5 of ‘funding’ in an envelope. She tells them that once they open the envelope, they have 2 hours to make as much money as they can.She cites three teams’ approaches:

1) First team opens a free stand that offers to check peoples bike tire pressure for free, then charges $1 to inflate if necessary. This team changes midstream to accepting donations instead of charging, and makes more money. Lauded for rapid iteration.

2) Second team makes lots of reservations at local restaurants, and then sells them to people waiting in line for same restaurant. Didn’t use the $5 at all. Lauded for realizing that the $5 constraint was artificial, and that using it constrained their thinking.

3) Third team skipped the exercise, and sold their 3 minute class presentation time as a advertisement to a local company. Made the most money. Instead of presenting, they recruited. Lauded for realizing that the 2 hours was also artificially hampering their thinking.

Personally, I found the presentation glib and condescending, and the solutions immoral and dishonest. I’m hoping for good discussion about the issues here, but I would not recommend the video.

I believe he thinks the solutions were immoral and dishonest because the first team initially charged a dollar for their services, when there was a free pump around the corner. However, that’s like saying Starbucks is immoral because they charge $4 dollars for milk and sugar. You can always go to the store, buy the ingredients and do it yourself.

Julian

1. Do you really need to create another “users” database or can you just use some other service’s vast db of friend-to-friend connections. Think facebook connect.
2. Do you even need to have your users login to you use your services. Why can’t it be anonymous.
3. This is no brainer. Use frameworks! Such as CodeIgniter for Php, Django for Python, Rails for Ruby, jQuery for Javascript… If you want to develop an app quickly and you’re idea is not so crazy that it can’t use a framework, why would you roll your own.
4. Relating to number 3, make sure you learn your tools well, because you might end up coding something up that has already been done, and that would be a waste of your time and money.
5. Use a css framework, for example Blueprint, to help you quickly create a grid based layout and try to use they’re classes as often as possible, while trying to avoid creating you’re own, as much as possible. Sure this limits you in certain things you can do, but it sure will keep it simple.
6. Take some time and see if anyone else has already developed what you’re thinking of rolling on your own. The amount of the you search until you find your solution, will be much less then actually creating it, maintaining it, debugging it, testing it and so on. Reuse code.
7. Don’t try to boil the ocean. If you’re idea involves solving multiple problems then be careful. The saying: jack of all trades, master of none wasn’t said for no reason. Do one thing and do it so well that people will beg you to use your site.
8. You really don’t need to have a blog for your startup in the beginning. Set one up after you finish your product.
9. Your startup idea can maybe be morphed a little to use another very successful site’s API. Then the people that are using that service might very well like what you’re building. For example, building an app on top of Twitter or Flickr.
10. Avoid creating complex relational database schemes when a less normalized version could do the trick. Or try out one of these newfangled schema-less free DB solutions: couchDB or TokyoCabinet

Now get to work!

Julian

FYI this post is in regards to startups/projects-in-general.

I’ve never started a successful company so take my startup advice with a grain of salt. However, I have tried a few different things over the years and gained some invaluable experience. My goal is to share what I’ve learned in hopes that it will cause you to take a step back and go, snap son, I’m dealing with the same thing! We’re all different people though and we work in different ways, so I don’t know, what doesn’t work for me, might really work for you.

If you’re like me, you get a little excited about certain startup/project ideas and it starts infinite looping in your mind to the extent that you have to tell every single person you meet your great idea. Don’t get me wrong, it’s not bad to be passionate about something. On the contrary, the ambition drives you to work hard and execute your plan. The problem stems from the fact that as the days go by, a routine is formed and set into place. Because its difficult to not take the path of least resistance, the routine goes on. By that, I mean that it’s more comfortable to do what you know will work.

You go to Starbucks and you think to yourself, hmm, well I always get a cup of coffee with half&half and two packets of sugar-in-the-raw because thats what I’ve always done. Maybe I should try this White Chocolate Mocha thing. I dunno though, it sounds kinda of whack, but then you buy it, take a sip and it’s like a kick-in-the-teeth good! You step out of your comfort zone, take a risk and discover something amazing.

Analogously, you go into work everyday and you sit in your cubicle or you go to your favorite coffee shop and bust out your Macbook Pro and begin to hack away at your idea. This goes on for quite sometime, maybe week after week, month after month and you get a boat load of stuff done. But just like how a frog doesn’t feel its being boiled as the water is slowly being turned up, so it is with you. You’re blind to the drop in your productivity and the stifling of your creativity. What you need is a White Chocolate Mocha my friend. You need to take a deep breath of fresh air. You need a slap in the face. I dare you to not work on your project, but rather just go and take a really long walk somewhere you’ve never been before. Go take some pictures and put them up on flickr. Get an Arduino and make it play Mary had a little lamb. As you start to do these things, watch how your brain starts to subconsciously think about your projects in all these crazy new ways. And guess what. That day or two days that you took off to just do whatever have payed off five fold because you were able to look at your problems in a new light, from a different perspective, and solve them in two days rather than ten days.

I just got a brand-spankin’ new Arduino Duemilanove from
Adafruit Industries and I just had to share the excitement. I downloaded their IDE and installed a simple program that pulse width modulates pin 9 making the LED fade in and out. The board can be powered either from the usb, the power jack or an external battery, which the board automatically selects depending on which one’s plugged in. I’m quite a fan of just having to plug in one cable, that being the usb, for power and data transfer. Thanks to their from-the-factory installed bootloader, uploading new code is a piece of cake. It takes care of copying code directly into flash and executing it when the device fires up. Simplicity is beautiful isn’t it? Here are some specs taken from the arduino.cc site:

To create the invocation:

Where t is of type (id) and s is of type (SEL).


NSMethodSignature * sig = [[t class] instanceMethodSignatureForSelector:s];
invocation = [NSInvocation invocationWithMethodSignature:sig];
[invocation setTarget:t];
[invocation setSelector:s];
[invocation retain];


Then to invoke the function:


[invocation invoke];


Here is the function I was trying to invoke:


- (void)_someFunc:(CocosNode *)node data:(void*)someData;


Now, the code NSInvocation code above is wrapped up in a class
and you have to pass in (id) t, (SEL) s and (void*)data. Then it takes
that info and it creates the invocation above. So I passed in:


[CallFuncND actionWithTarget:self selector:@selector(_someFunc) data:someData]


See the mistake? The selector I passed in was just _someFunc instead of _someFunc:data:
and so when it created in the method signature it wasn’t able to find the method and so
it returned nil, which was being passed into NSInvocation and so it seg faulted. This happened
because I’m still used to C and C++ where you a function is defined as int _funcName(param1, param2). So the fixed code is actually:


[CallFuncND actionWithTarget:self selector:@selector(_someFunc:data:) data:someData]


Here’s a video of it in action. Some random guy in Japan made the video, I found it by searching for conactSwap on google.

Long story short, I unlinked it and everything stopped working. No more ls, mv, ln. Nothing. So after frantically searching on google I stumbled upon this:

> export LD_PRELOAD=/lib/libc.so.5.0.9 if you’re using bash,
> setenv LD_PRELOAD /lib/libc.so.5.0.9 for tcsh

You have to point the variable to a valid libc library and it’ll temporarily fix the prob. After I did this, I linked it back to the original libc and everything worked perfectly. The funny thing is, I thought there was a newer version in /usr/lib, but it turned out that it wasn’t valid. So I just pulled the hair out of my head for no reason while learning to never unlink a library ever again.

I’m such a n00b.