Re: Comments on Math_BaseX - was Re: [PEAR-DEV] Question about package.xml

On Wed, Dec 11, 2002 at 03:40:59PM -0800, Jesus M. Castagnetto wrote: > Yep, I gathered that much from the code. What threw me off was that it was > included in the Math_* namespace, which is why at first approach I was > expecting to conform to base standard used in math where for the first 36 bases > you use the alphanumeric charset, and from then on you are on your own ;-) > > Perhaps you can do both and support the normal mathematicl base char sets or > random charsets. For now it works like the way i intent it to work and is ready for implementing. Support for strict mathematical base is maybe an option for version 2. > For example, the method that initializes the valid set of chars you could > accept 2 parameters, check if the first one is numeric or a string. If it is a > string, then the user wants to use a non-standard charset base (like in your > case), and then ignore the second parameter. If the first parameter is numeric, > check if the second parameter is a string, which will mean the same as the > previous case, with the addition that you can also check that the base and the > number of chars in the string match. Finally, if the second parameter is not a > string, ignore it and use the usual mathematical conventions, just making sure > that 2 <= base <= 36. Did you knew i had to read this part at least 3 times before i understood what you're we're saying. Drawing a diagram helpt me a lot ;-) But seriously, it's sounds good. I'll keep it in mind designing the new version. > Additionally, when a non-standard charset is being use, there should be a > method that checks for uniqueness of all chars (count_chars() will here here). This function is already inside the class. It loops the character string and does a substr_count for each char. It a char occurs more than once, an pear error is raised. There are also checks if the code passed to toDecimal function are allowed, is the base is set (Without a base, a conversion is impossible) and so.. > > I use this class to encode an ordernumber for use in an email. The user must > > reply by visiting the link in the email. > > We're using base13 (Only 13 characters from the alfabet). > > > > This class is indeed basicly doing the same as base_convert only with a > > variable character set, > > I see. That is more akin to data encoding that numerical base transformation > which has a definite meaning in math. Text_BaseEncoding might have been a more > explicit moniker perhaps? That's my my first methods were encode and decode. But toBase and toDecimal is also working for me But by passing a hex character set a strict mathematical conversion is also possible. And because it based on the mathematical algorithm as the base conversion I categorized it as Math. No-one objected so far on the namespace, and the next version will also do the normal mathematical conversions. But still advise people to use the normal PHP base_convert function in stead of using my class.. Tomorrow i will release the current package and tag the current branch. Than i can try to implement the PHP base_convert for strict mathematical conversion. The easiest way is just the substr (base length - 1) the part nessecary of '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ", but than i would bypassing the base_convert function. Something is wish to prevent. So i first have to work these things out before i even start coding it ;-) -- With best regards, Dave Mertens - Senior Development Manager dmertens@ism.nl Innovative Solutions in Media Schiekade 101 3033 BG Rotterdam, Netherlands T +31-10-2436060 F +31-10-2436066 http://www.ism.nl Quality Solutions - Reliable Partner