In the beginning. ...

After 35 years of working and playing with language creators and improvers, I decided to try my hand at a language of my own, one just for my own amusement.  Of course, being me, I immediately went off with grandiose plans, far more things than I could realistically accomplish.  After a few bouts of reality, got down to the core of what I wanted:
     1. A language without vowels, with the syllabic peaks being always consonantal continuants and no phonemes having allophones in the vocalic triangle (in the paradigm dialect alt least).  This was the amusing parts, since I like the sounds that are produced, including the language name, pft (subject to change as the program progresses)
     2.  A reasonably natural language (phonology aside), with quirks and peculiarities scattered here and there, but, perversely, in regular ways.
     3. Since, at the beginning at least, the temptation for a language creator is to make the language regular across the board, I needed some way to build in the irregularities I wanted.  Since most of the oddities of most languages are a product their histories: phonological shifts, borrowings, underlying languages of conquered people (or conquering), I decide that my language must have a history, first language from which it was descended through the various vicissitudes of language development.  That is, I will build my contemporary by building a nice orderly ancient language, then screwing it up in the ways languages have managed to screw up their ancestors (or, rather, the inverses of all the ways linguists have devised to get nice orderly languages from the muddle of their descendents).  I am already beginning to see that thre fun is going to be in the process, not so much in the product.

Since all my non-field-related linguistics is in Indo-European languages, my pattern for the ancient language is UrIndoGermanisch: roots strings of consonants, surrounded and separated by vowels which pop into place out of somewhere, under the influence of accent, phonetic environment and affixes, broad ranges of cases, many verbal forms, prepositions (maybe post- for variety), derivatives and combination rules.  I expect I'll begin with something a little simpler than UIG, though. in all these things.  The move then will be by staggered waves of phonetic shifts (vowel weakening, palatalization, rhotacism, ...) and grammatical shifts (periphrastic constructions, word order fixing, reanalysis, ...) and environmental influences (meaning changes with culture, borrowings and incompetents, ...)  I will try to abide by the two great (contradictory) rules of the Junggrammatikern "No exceptions to sound change rules" and "Every word has its own history."

Needless to say, all this planning is to avoid actually starting working and, especially, starting to think of vocabulary.  I'll get my start on that from the usual source, what others have used as basic -- probably (because I know them best) aUI, NSM, toki pona and Swadesh -- plus my own more or less a priori ideas of what is basic (partly derived, however, from problems with the above cases). I don't yet have an idea how to assign roots to various concepts (except that the root for "speak" is ppd), but, if nothing occurs to me, there is always randomness.