Gengen 1.1

With great pride I now present the latest version of Gengen! Version 1.1 has seen critical improvements to the way names are generated, resulting in huge strides toward the end goal of "every name is good, every language is good". Here are the highlights:

• Name length is much more carefully regulated, resulting in fewer overly long names like "Gapyetjachmubhong" and overly short ones like "U". Moreover, name length is now tied to the robustness of a language's phonemic inventory and the permissiveness of its phonotactics. What this means is that languages with simpler syllables tend toward longer names, while languages with complex syllables can attain the same degree of distinctiveness with shorter ones. Now it's more likely to stop  generating at "Gapyet" or "Gapyetjach".
• Stress assignment has been totally overhauled and made way more transparent. Primary is now marked with an acute accent if it falls in an "unexpected" location, and all stresses are marked explicitly in names' IPA representations.
• Consonant inventories have been revised to include a degree of natural irregularity, making them appear far less contrived. Additionally, the voiceless alveolar sibilant affricate ⟨t͡s⟩ now has a chance of appearing under the representation 'c', like you might see in Pinyin and certain Slavic orthographies.

The web interface has also been updated to highlight these changes, provide more information, and make the Gengen experience more user friendly:

• If you're not sure how to pronounce a language's names, you can now reveal a suggested pronunciation for every name with the "Show IPA" button. If you don't read IPA, you'll have to wait a little longer for us to add phonetic pronunciations.
• You can order a list of names alphabetically, but the new default is ordering by "complexity", a measure of how unlikely a name is for its language. The advantage of this ordering is that most unrealistic names tend to have very low or very high complexity and wind up at either end of the list. If you're curious, you can toggle names' complexity measurements with the "Show infocon" button - the best names tend to lie between 6 and 20.
• A new "Language data" tab now provides additional information about a language. For now, it only describes syllable structure and displays the language's phonemic inventory, but more information will come in future updates.
• The "Random seed" button has been replaced with a "Random language" button, which instantly generates a new language when clicked.

With these additions, Gengen is a lot closer to our original vision, but some languages still fall short of our standards. Luckily, we've got a lot of exciting solutions in the works. In addition to simplifying the structures that produce names from phonologies (allowing for easier fine-tuning), we're looking forward to adding geminate consonants (like "gg" and "rr"), characteristic initial letters (English loves a "J" at the start of a name, not so much in the middle), and suffixes (one of the simplest ways to distinguish languages), and maybe cleaning up some of those unsightly, unreadable sequences of "Y"s that love to pop up in certain languages.

Thanks for your continued interest! Big things are coming - keep watching this space.