Sophie: tsid-0.6-1mdk i586

tsid-0.6-1mdk.i586.rpm

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
   <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
   <meta name="GENERATOR" content="Mozilla/4.75 [en] (X11; U; Linux 2.2.16-22 i686) [Netscape]">
   <meta name="Author" content="Tognon Stefano">
   <meta name="Description" content="A Faq about TSID">
   <meta name="Keywords" content="TSID, SID">
   <title>TSID FAQ</title>
</head>
<body text="#000000" bgcolor="#FFFFFF" link="#0000EE" vlink="#551A8B" alink="#FF0000">
<b>General FAQ about TSID</b>
<p><b>Q: What is the porpoise of TSID?</b>
<br>A: TSID (Time Sid Manager) is essentially a way to collect your TSL
(Time Spent Listening) statistic about your listened sids, and consequently
it may elaborate these statistic to found some great information.
<p><b>Q: So TSID is a new sid player?</b>
<br>A: No, it is not a player. It's a library that a player may use to
collect information.
<p><b>Q: What kind of information can be extracted from my stored statistics?</b>
<br>A. Well, the program <i>examiner</i> can produce sorted list of your
listen tunes. For examples it may produce a sorted list of tunes you have
listen for more time, or the authors you listen more often.
<p><b>Q: <i>Examiner</i> has a pondered option. What is this option?</b>
<br>A: TSID stores two values for each tune you listen: the total second
you listen it, and the number of time you listen it. These information
are used for producing ordered list. However if I&nbsp;want to try to determine
my best sid/composer, I must use each values for ordering the list, using
an algorithms that merge the two value and produce a score.
<p><b>Q: How is calculated the score in pondered option?</b>
<br>A: The idea is this: if you listen to a tune for his length, you optain 100
points (100% of a tune); a 50% of length produce 50 points. This looks good for
all tunes, but not for looped tune. If you listen a short repeated tune, you can
obtain a big point. So, a interpolation rate is used to limitate this effect.
However, it's probably that the paramethers used may changed in the future
for reaching a better result.
<p><b>Q: May I use TSID for collecting information about sids not in HVSC?</b>
<br>A: This is a good question, because TSID was developed with HVSC in
mind. Actually TSID put the information in it's directory looking at the
passed file names, so a HVSC file may be passed like: Gray_Matt/Driller.sid.
But if I pass a file like My_Sid/Author_x/tune1.sid, this directory will
be created in the main TSID directory. If "My_Sid" is different between
all HVSC directory, you may collect information to the other your collection
without having problem with HVSC.
<br>
I'm afraid if TSID should support other collections, because HVSC is the most
complete available. So, if you want to use other collection, you may put the
others your collection in a new directory inside HVSC, this is not philosophically
correct, but works.
<p><b>Q: I see that tunes listen for less than 10 seconds are not stored.
Is this a bug?</b>
<br>A: No. This is a choice to avoid to include the short sound effects and for
allowing a user to choose the tune he wants, instead of the default starting up
when open a sid file.
<p><b>Q: I had used a sidplayer with tsid but I didn't apply the update file to
the stored information after each HVSC upgrade. Can I apply all the updates now?</b>
<br>A: Yes, but there are some things to known. The update process moves the
stored information as the new upgrade in HVSC require. So if a file was
/DEMOS/Pippo.sid and now it is /best/pippo.sid, the update process takes the
information stored for /DEMO/pippo.sidt and move these to /best/pippo.sidt.
If /best/pippo.sid already exists, the information are merged, so you can apply
the update after having listen to some HVSC upgrade. <br>
However there is a case that this can not be done: if some subsongs changes their
positions from an upgrade. At this point if you don't apply the tsid update, when
you listen that subtunes, the stored information goes onto the wrong positions.<br>
Hohever my experience says these: in one HVSC upgrade only 3 tunes are of these kind,
so the probability to listen to them with more than 16.000 sid is little.
<br>

<p><b>Tognon Stefano FAQ about TSID</b>
<p><b>Q: Why do you call the project TSID?</b>
<br>Because TSID sound good. Just a pronunciation choice.
<p><b>Q: Why&nbsp; do you have thought to realize this project?</b>
<br>A: In that days, I was looking at top 10/100 sid tunes lists that ca
be found in the net. These lists are the result of people voting about
their best tunes. Why vote? The best way should be that my sid player give
me my list automatically looking at my listen history.
<p><b>Q: When did you decide to realize the program?</b>
<br>A: In year 2000 I was at military service (as conscientious objector)
and this occupation took&nbsp; me all energy. At September I had realize
that I loosed 90% of my programming abilities during this service, so I
wanted to start with a C++ project that give me the opportunity to restart
to program.
<br>If you look at the code, you will see a lot use of pointers: just a
way to not forget them.
<p><b>Q: I don't think that my time statistic will say who is my best sids,
so why collect these information?</b>
<br>A: Just for curiosity: how many sids had you listen until now? 5%,
10%,... 100% of HVSC?
<br>But there is a hidden features of TSID that I want to illustrate. When
I looked at the HVSC the first time (late 1999), in a month I listen at
random many sids, just like to give an idea about the collection. One day
I listen to a sid and after a week I realize that this sid was a cover
of a french song. So I try to remember where the sid was in the HVSC looking
at the sid player history, but unfortunately the tune was fall out of history.
<br>The last hope: STIL. No, the cover is not already in STIL. So I'm still
looking at that sid without result.
<br>But if I were using TSID, I can found the sid in a simple way: looking
for files stored by TSID that were last modified in the date I listen the
tune. If I remember the date correctly (1 or 2 days of errors) this can
reduce to search to a 10/20 sids in the best case (or 100/200 in the worst,
but this is better that looking at 17000 sids).
<br>
</body>
</html>