[Snowball-discuss] Problems with step 5 in the Porter2 algorithm

Håvard Lindset lindset@webpixels.net
Sat Oct 18 19:50:01 2003

Thanks for the reply, Martin :)

So if the e is in R2, the first consonant must be there, while if it's in
R1, it doesn't have to?

So if it's in R1: [C]VCe
And if it's in R2: CVCe

Or did I misunderstand what you meant?

Thanks again :)

Best regards, 

-----Original Message-----
From: Martin Porter [mailto:martin.porter@grapeshot.co.uk] 
Sent: 18. oktober 2003 17:15
To: Havard Lindset; snowball-discuss@lists.tartarus.org
Subject: Re: [Snowball-discuss] Problems with step 5 in the Porter2

>Should I check just DIRECTLY in front of the ending e, or shouldn't 
>there be ANY short syllables at all in the word before the ending e?


Check DIRECTLY before the e. If e is in R1, but not in R2, you must have a
pattern [C]VCe, where C is a list of consonants; V a list of vowels. [C]
means the list of consonants may be absent. [C]VC is short if V is one
vowel, and the final C one consonant other than w,x,Y. So you don't remove
the e from ape, tape, scrape, but you do from able, table, whistle, house,
goose, grieve.