In some cases, you can also use your database engine to generate a GUID. In
MS SQL, put this in a DirectDB action:

SELECT NEWID()

@@resultset$[1,1] will be your GUID.

Bill





On Thu, Oct 14, 2010 at 3:38 PM, Robert Garcia <[email protected]> wrote:

> This is a weakness in witango, but easily shored. Witango lacks the ability
> to generate a GUID, but would probably be an easy feature to add, since I
> think it can be done by the host OS.
>
> There are a few ways to shore this in witango...
>
> 1. If windows witango, find an old but reliable tool, called GUIDMKR. Its a
> com object I have used for YEARS with witango. Let me know if you need code.
>
> 2. In mac or linux, use a command line, this can be done in pearl.
>
> You could then have code in witango turn that into your GUID into a BYTE
> array, so it looks like a code, like a YOUTUBE id, much shorter.
>
> http://stackoverflow.com/questions/2983065/guid-to-bytearray
>
> <http://stackoverflow.com/questions/2983065/guid-to-bytearray>I worked on
> this for many hours many moons ago. There was always dupes, the better my
> algorithm got, the less dupes, but to be safe, start with a TRUE UUID or
> GUID generator. GUIDMKR has always served me well on windows/witango.
>
> --
>
> Robert Garcia
> President - BigHead Technology
> VP Application Development - eventpix.com
> 15520 Coutelenc Rd
> Magalia, Ca 95954
> ph: 530.645.4040 x222 fax: 530.645.4040
> [email protected] - [email protected]
> http://bighead.net/ - http://eventpix.com/
>
> On Oct 14, 2010, at 12:18 PM, Roland Dumas wrote:
>
> I need to stare at this code a bit more. I made it witango 5 compatible
> (while object and instead of @sleep, I just threw away output based on the
> same timestamp)
> results
>
> code: NZVWIXXX timestamp: 1287058511
> code: LURAVSXX timestamp: 1287058512
> code: LURAVSXX timestamp: 1287058513
> code: NZVWIXXX timestamp: 1287058514
> code: ZLNKFQXX timestamp: 1287058515
> code: BNCILFXX timestamp: 1287058516
> code: KRCCOCXX timestamp: 1287058517
> code: NZVWIXXX timestamp: 1287058518
> code: BNCILFXX timestamp: 1287058519
> code: BNCILFXX timestamp: 1287058520
> code: NZVWIYXX timestamp: 1287058521
> code: KRCCOCXX timestamp: 1287058522
>
> Interesting, don't you think?
> In a long sequence, the repeats are frequent and and trailing Xs...
>
> Thanks for the start.
>
>
> On Oct 14, 2010, at 10:53 AM, Robert Shubert wrote:
>
> Try this out:
>
>
> <@! start with epoch>
>
> <@ASSIGN secs <@TSTOSECS <@CURRENTTIMESTAMP>>>
>
>
> <@! add salt>
>
> <@ASSIGN secs "<@RANDOM low=1 high=9>@@secs<@RANDOM low=1 high=9>">
>
>
> <@! convert to base 26>
>
> <@WHILE expr="@@secs > 26">
>
>
> <@! shift to ascii for capitals 65-90 convert to char and build code>
>
> <@IF "<@CURROW>%2 = 0">
>
>  <@ASSIGN code '<@CHAR <@CALC "abs(@@secs % 26) + 65">>@@code'>
>
> <@ELSE>
>
>  <@ASSIGN code '@@code<@CHAR <@CALC "abs(@@secs % 26) + 65">>'>
>
> </@IF>
>
>
> <@! shift place for base conversion>
>
> <@ASSIGN secs <@CALC "@@secs / 26" precision=0>>
>
>
> </@WHILE>
>
>
> Final Code: @@code
>
>
> DISCLAIMER: It is possible to create the same code if this routine is run
> multiple times in the same second. Add a <@SLEEP 1000> to prevent this.
> Also, there is an abs() in there because doing modulo on large values
> produces a negative value. This is incorrect behavior on Witango's part and
> something that I'll need to fix in the server. I do not think it produces a
> collision problem in this routine, but it may. It is likely tied to small
> floating point variable used in the routine. Oh, and there is also a
> rounding error that I didn't work out, but again, it only changed some shift
> points (26^1 place increases at unit=13 rather than unit=0). This can
> removed by floor()ing the base conversion shift in the final @CALC. Again, I
> do not believe that this causes a collision situation. Also, this routine
> can occasionally create a 7 character code. You may need to check for and
> ignore these if you want them all to 8 characters (This may be able to be
> eliminated by increasing the LOW value in the first @RANDOM) Lastly, I don't
> guarantee this routine. I just whipped it up as a mental exercise, so you
> may want to test it out some (see if you get a collision over a few million
> iterations), or show it to someone who is good with algorithms or
> cryptography.
>
>
> Robert
>
>
> -----Original Message-----
>
> From: Roland Dumas [mailto:[email protected]]
>
> Sent: Thursday, October 14, 2010 11:42 AM
>
> To: [email protected]
>
> Subject: Witango-Talk: generating unique codes
>
>
>
> looking for method to create unique and non-obvious n-digit codes to be
> used as discount codes.
>
> Should be short enough to use as consumer codes (8 digit?), and probably be
> something simply to enter, such as an all caps, and not have an easily
> discoverable pattern.
>
>
> I want to generate unique codes and not have to check the history to see if
> a code has been used before.
>
>
> Suggestions?
>
>
>
>
>
>
> ----------------------------------------
>
>
> To unsubscribe from this list, please send an email to
> [email protected] with "unsubscribe witango-talk" in the body.
>
>
>
>
> ----------------------------------------
>
>
> To unsubscribe from this list, please send an email to
> [email protected] with "unsubscribe witango-talk" in the body.
>
>
>
>
>
> ----------------------------------------
>
> To unsubscribe from this list, please send an email to
> [email protected] with "unsubscribe witango-talk" in the body.
>
>
>
> ------------------------------
> To unsubscribe from this list, please send an email to
> [email protected] with "unsubscribe witango-talk" in the body.
>


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

To unsubscribe from this list, please send an email to [email protected] 
with "unsubscribe witango-talk" in the body.

Reply via email to