A program to compute number of days between two dates?

mrclisdue · 03-16-2013, 09:23 PM

Quote:

Originally Posted by psionl0

Code:

|                                    *
+
|
|
|
|
+                  ----------------------
|                     *       *
|                      *
|                                   *
|
+                               *
|                       *  *   *   *
|           *                     *
|               *    *
|                           **   *
+                 *---------------------
|      * *                *
|                   *
|  * *           *
|*  *      *  **
+ *       *
|
|     *
|
|
+------------------*--------------------

I'm sure I can see Bob Dobbs, and perhaps a pipe.

Tho' it could also be a cat, moog synthesizer, or a baby's arm holding an apple.

cheers,

w1k0 · 03-16-2013, 09:28 PM

Quote:

Originally Posted by qweasd

I don't think "release" can be put on a real line without taking serious liberties with interpretation, which is why I called it voodoo. My X axis is not a mystery: it's just 13.0 for ver. 13.0, 13.1 for ver. 13.1, etc.

I don’t try to argue with you. I like very much the name “voodoo” that you chose for that type of the the data. When after 1 goes 1.1, and 1.1.2, and then 2.0, but after 4 goes 7, and after 13 goes 13.1, and 13.37, and then 14 it’s complete voodoo. The English language isn’t my native one. Moreover I never learned English language. So I thought “voodoo” is the standard name used in the English language in such a case. Now I know you invented it and I admire that name much more than before. I simply like very much the people which are able to invent something.

Quote:

Originally Posted by qweasd

But assigning it this way is just as arbitrary as mapping each release to an integer, or whatever is that you did...

It isn’t whatever. See the descriptions of the axes. X is for the days after Slackware 1.0 and Y is for the days after the previous Slackware release. So both axes display the days. I tried to use the same scale for both axes but the resulting graph was very flat so I gave up and I used the standard scale.

Quote:

Originally Posted by qweasd

After all, the version assignment is a random experiment from hell; it is dependent on everything, including this thread. What is your first column anyway?

The first column (X axis): the days since 1993-07-17 (Slackware 1.0 release)*.

The latter column (Y axis): the days since the previous release.

* There’s 121 days till Slackware’s 20th birthday. Let’s celebrate that day when it’ll come!

Quote:

Originally Posted by qweasd

And using trend here is also questionable: the correlation coefficient is pretty low, the spread is huge, and so it is far from certain that the mean went up. But for purely gambling purposes I would bet on the trend, probably because I always found pictures very convincing

I didn’t plot any trend: qweasd plotted the simple linear regression and allend plotted the regression slope and some standard errors. I don’t consider these plots reasonable or useful.

w1k0 · 03-16-2013, 10:53 PM

Here’s the official graph of the slacky randomized standard deviation coefficient divided by some imaginary number that I imagined almost one hour ago:

Code:

+             **************             
|           **              ****         
|         **                    *        
|       **                       *       
|      *                         *       
+      * ****              ****** ***    
|     *       *************       *  *   
|     *   *                        *  *  
|     *   *                        * *   
|     *   *                        * *   
+     *   *                        * *   
|     *   * ****            ***** *  *   
|     *  *      **        **      *  *   
|     * *  ********     **         * *   
|     ** *    **  **   **  **      * *   
+     ** *    ****        *        *  *  
|     *                     *      *  *  
|     *            *               *  *  
|         *      **      *          *    
|         ****  ***     *   *   **       
+         ***   *******      ****        
|         **      ***    ****    *       
|         * *   *         *     *        
|          * *  ***      *     *         
|              *****          *          
+            **              *           
|    *  *****               *            
|  ***  ****    *         *              
|   ******         *** ***               
|     ***                         by w1k0
+----+----+----+----+----+----+----+----+

The function used to draw that graph is too complicated to be stored on all LinuxQuestions.org servers – I’m sorry!

For mrclisdue, qweasd, and the other guys (male sex) and gals (female sex) as well as for all the asexual creatures (neuter sex) in the whole universe and beyond with the special greetings for the exceptional human being living here and now and named volkerdi (male sex).

w1k0 · 03-16-2013, 11:29 PM

Quote:

Originally Posted by mrclisdue

I'm sure I can see Bob Dobbs, and perhaps a pipe.

It seems that the next points on that graph will be the smoke from the pipe: that gives the valuable clue for all the forecasters (it’s enough to monitor the movement of the air molecules and the changes of the temperature to print the next point in advance).

psionl0 · 03-16-2013, 11:38 PM

Quote:

Originally Posted by mrclisdue

I'm sure I can see Bob Dobbs, and perhaps a pipe.

I guess stargazing is as good as crystal ball gazing.

w1k0 · 03-17-2013, 03:04 AM

Quote:

Originally Posted by mrclisdue

I'm sure I can see Bob Dobbs, and perhaps a pipe.

I have to argue with you a bit because I – on the contrary – am sure that I see the pipe, and perhaps Bob Dobbs.

No offense...

w1k0 · 03-17-2013, 03:39 AM

Some of you may be curious which function I used to draw the nice graph from the post #63. As I mentioned above that function is too complicated to be stored on all LQ machines but I can disclose the shortened version of that function:

Code:

y = f(me)

Unfortunately the current version of gnuplot doesn’t understand that simplified function so you can’t re-draw my graph alone. (Feel free to send the bug report to the gnuplot developers.)

w1k0 · 03-17-2013, 06:38 PM

The data file and the graph from my post #59 show that during each consecutive 1000 days there appeared less and less Slackware Linux releases:

Code:

    0–999: 8
1000–1999: 6
2000–2999: 5
3000–3999: 4
4000–4999: 3
5000–5999: 4
6000–6999: 2
7000–7999: ?

That means that Slackware is very mature Linux distribution now (it ended 18 years three months after 13.37 release).

Didier Spaier · 03-18-2013, 12:37 AM

There are three kinds of lies: small lies, big lies, and statistics.

eloi · 03-18-2013, 04:35 AM

Quote:

Originally Posted by Didier Spaier

There are three kinds of lies: small lies, big lies, and statistics.

I remember a young friend, he was very keen on science. He felt insulted when I discredited statistics. Be careful, don't criticize idols in front of religious people.

w1k0 · 03-18-2013, 10:13 AM

Definition of a statistician: Someone who, if you put his feet in the oven and his head in the freezer, would say, “On the average, I feel just fine”.

schneidz · 03-18-2013, 12:42 PM

Quote:

Originally Posted by colucix

You don't really need a program. In bash you can try something like:

Code:

echo $(( ($(date -ud 20130331 +%s) - $(date -ud 20130314 +%s))/86400 ))

The code might be transformed to a shell function that accepts the two dates as arguments. Just my

.

^ +1, my stab at it:

Code:

expr `date -d 20130316 +%Y%j` - `date -d 20130319 +%Y%j`
-3

Kallaste · 04-26-2013, 10:13 PM

Quote:

Originally Posted by w1k0

That’s rather natural idea in such a case. To “predict” the date one can use different methods. The more methods she’ll try the more results she’ll gain. As a result one of these “predictions” will be really close to the actual date of the next release. That’s the whole truth about the predictions and the numerology.

Ambiguity in predictions and a continuous supply of new data in which to look for coincidental similarities doesn't hurt, either. I see it as the Nostradamus effect. Given enough time and enough room for interpretation, any prediction will "come true."