How should names be romanized which have an "alef" acting as a vowel and no clue as  to whether it's an "a"  or  an "o"?
Do  we choose a form, make  a reference from  the alternative, and  then  code it as provisional?

The answer to  the second question is  "yes."

As  for  the  first question, the answer will vary with circumstances.  For example,  the  name  "shin-nun-alef-resh"
matches up  nicely with  a well-attested German name, Schnorr, so we can choose "Shnor" as  the entry element and  make refs from  (at  least) "Shnar"  and "Shene'or," and  from  other possibilities if  you think of  them.

Note:  With a question like this, it's  a  good idea to give the whole name and maybe the name of  the  book that provoked the   question.  This information might  help someone on  the  list  to locate a more specific answer in a reference source.

Alef at the end of names of Slavic derivation

Certain names of places and people of Slavic derivation, that normally end in yod, are occasionally written  with  final  yod-alef.   The  most  common place-name like this is probably Bilgorai, written
           bet-yod-lamed-gimel-alef-resh-yod-yod
           bet-yod-lamed-gimel-vav-resh-yod-yod-alef
           bet-lamed-gimel-resh-yod-yod-alef
and probably other ways too.
Similarly, Joseph Zevi Dushchinsky whose surname was spelled
           dalet-vav-shin-yod-nun-samekh-kuf-yod-alef

Do  these  spellings  represent  variant pronunciations  (Bilgoraya?  Dushinskaya?)  that should be recognized in romanization?  The answer is that in cases like  Bilgorai, the  alef should be regarded as  a vestigial spelling convention, and not as an indication of pronunciation, and be ignored in romanization (i.e., Bilgorai  and not Bilgoraya)

Romanizing 'Ayin and Alef in Abbreviations

If  "'ayin" is the *first* letter of the word being abbreviated, it has to  have  a  vowel  following  it.   It's  okay  for  a  word  or
abbreviation to *end* in  an "'ayin".  Similarly, an  "alef" which is the first letter of an abbreviation has to have a vowel following
it, but since we don't write the initial "alef" in these cases, the similarity to  the situation with "'ayin" is  not obvious.  The
problem  arises  when   "alef"   is   the  *last*  letter  of   the abbreviation, as  in "ma'." for "ma'amre."  When a *word* ends in
an "alef", of course, we  also don't write it,  but just   writing   "ma."   would   not  adequately  represent   the abbreviation in  this case--and besides it  would cause people to say, "But I thought abbreviations couldn't end in vowels?"

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

When "Venetsyah" is spelled  with  an  alef (i.e.  it  ends yod-alef-he) on  the t.p., this does not mean that  it should be romanized "Venetsi'ah".  If the alef occurs in  the vicinity of  a patah or kamets, one can  treat it as an  "em keri'ah".