Understanding databasification is a tricky endeavor. Hiroki's language makes it seem like he's against linearity or that linearity is antithetical to flat, postmodern readings. I can understand a lot of critique levied against him. Regardless, I think the fairest way of understanding databasification is this amorphous 'soup' whereupon pieces are drawn and constructed.
In fact, while I'm being a bit reductive in an argument that heavily draws from Derrida, Kojeve, Deleuze, Lacan, and Zizek, I think that if you engage in the very act of "oh this is another moe-blob" or "oh this is another ecchi" sort of argument you've been engaging in some level of database reading!
Where I disagree with Azuma is that readings themselves are valueless. Recall Eisenstein's famous apocryphal mantra: 1+1=3. Parts alongside parts can lead to greater than its surface sum. Lucky Star did it. Madoka did it. Anime and Manga are rife with value-laden narratives born from the postmodern soup.
Really, I think the short length of Otaku: Japan's Database Animals does it a disservice. Azuma's main argument uses - curiously enough - visual novels. Since visual novels can be parsed into separate components, they are, quite literally, deep layer narratives.
However, at the heart of databasification is technology. Azuma references the internet and visual novels; Manovich references a variety of new media. This argument has roots as far back as Deleuze. If we abstract the topic even further we can find roots to the Soviet school. Ultimately, at the heart of databasification - and curiously Netoyome - is one of many theories on the phenomenology of storytelling juxtaposed with the development of technology.