I really like this one.

What happened, Aleksandr, going for a kind of hexameter? Abandoning the linnod? ;-)

+Tamas Ferencz Is something wrong? Not quantitative, indeed, this time I didn't manage it, but I thought the accentual pattern of the linnod is followed? TÉR-bad-na-ÁN-nûn-GLÉN-nas, ÁU-rath-hín-RÁEG-i-BÁ-dath.

+Александр Запрягаев nope. No problem. Now that you mention it, I am not even sure how accentual pentameters work. It's there a max number of syllables in one half?

Maybe I misunderstand the syllable divisions in Sindarin, but I believe there are seven syllables in each half, long vowels and diphthongs counted as one syllable (except ui, which is allowed to be split in two); the attested one has the accentual measure —oo—o—o/—oo—o—o, which I always exploit, as well as a quantitative measure —oo—oo—/—oo—oo—, which is a classical (dactylic) pentameter and which I also try to retain, though succeed quite rarely, for Sindarin seems not really fit for such compositions, even most of the verbal forms cannot be used in them at all. Accentual pentameters are supposed to retain the structure of —oo—oo—/—oo—oo—, but in stresses instead of length, but linnyd do not use such a metre
I use Salo's interpretation when a syllable is counted 'long' if houses a long vowel, a diphthong, a cluster of two or more consonants after the vowel (even appearing across a word boundary, except ss, which is considered short) or a single consonant m; and short otherwise — quite a limiting system when you want the majority of your syllables short.

+Александр Запрягаев Do you know of any authentic example showing that an s written double doesn't make its syllable long?