While soubmitted my new pad file I've obviouly been examining it on softtester.
Although I do look at pad file addition submissions and convert them to update submissions if they are re-submitted. Somehow I've now got 3 copies of mdbsecure on softtester. One from my old site and 2 from the new.
Although the 2 new ones are using the same author record.
Joy, just wonder how many duplicates other authors have :(
by JM
Ahh I'm actually determining unique programs / pads incorrectly, by PAD URL and what I call PAD CRC, which is a CRC value of the string contents of the pad file.
ReplyDeleteSo obviously this is wrong as the reason a pad file gets re-submitted is that its been updated and therefore PAD CRC will be different.
So what makes a program unique?
I would, off the chuff, say its name, but names can change over time, like my two versions, including the year.
Suggestions?
I guess if we use my program / pad file as an example to talk about its a good starting point.
ReplyDeleteI think your point 2 is flawed as a company can have many products.
OK, let us try and come up with a definitive list of things which we need to check.
1) unique pad url in database, I guess this would solve my new mdbsecure problem straight away.
2) Unique download url in DB, again this would solve....
3) What about program names / pro version etc, I'm not sure if theres an easy way to solve this. I mean we could remove spaces, numbers, full stops and some keywords like PRO / LITE etc. But I don't think this would be definative.
I guess point 3 would mean I would only have 1 MDBSecure in my DB. Which I guess strictly speaking should be correct.
Thoughts?
Well I've fixed my problem, with duplicate pad urls, which is OK for now'ish.
ReplyDeleteI already had a delete duplicates php script from earlier this year, had forgotten. But I should get any more, done a program update.
However I think its a good idea if we discuss this further and come up with a near bulletproof solution.
Sure thing. I'll read my code more closely and double-check the duplicate detection algorithm.
ReplyDeleteI think point 2 in my firt comment is OK. It will only say a PAD is duplicate if both the company and program name are the same so a company can list many products without being deleted.