TLDR:
no, not a good idea. as it stands.
for one, it doesn´t serve the internal functions of any dedicated DAM
for two, it doesn´t allow for real/robust integration and roundtripping with other photo- and DAM-tools
– but try Photo Supreme, if you still want to buggle up your photo management w/ DT… in ‘some way’
–
… some additional thoughts:
I myself belong to the group tempted to loop-in DT into my photo/image/visual media workflow when it comes to creative forms of information and knowledge handling (sorting, finding, retrieving, associating etc.).
One thing is that one feels DT offers a kind of friendly invitation, as
- its media display is rather advanced (just think of annotating sub-parts of videos…)
- DT effectively does read (and makes searchable) some of the EXIF/IPTC-metadata (even though in the end this is all not very transparent in actual use, as the UI is not showing all of it)
- the whole power of aliases, finding related content/folders, and (now) automation
… so there is a real ‘seduction scenario’ here 
add to that on the more ‘meta-level’ that
- some people (like me) do think that information is semiotic and not textual in nature. for people like that/me the separation of text vs. image/graphic/schematic/etc. just doesn´t make sense in the 21st century. (even more information is one kind of beast; meaning-making resources another…)
- for some people probably there is a ‘natural urge’ to find one privileged (if not exclusive) place to channel all one´s knowledge assets + sense-making resources to and manage them together – and DT is just one of the best candidates…
then I would agree with Bluefrog, after many attempts to bend DT that it´s not really a ‘good idea’ (inspite of some of the ‘invitation signs’).
some of the larger stumbling blocks are:
- no display of nested files on higher folder levels (if you look at it all serious DAMs allow for that; and with visual material that is 'for a
- the bridge btw. EXIF/IPTC/XML metadata and DT-metadata handling is only partial and not (easily) obvious/transparent, probably fragmentary. (some IPTC metadata are read, findable; it seems of them are diplayed in DT-panels some not; then others are not read…etc.); also there is no way to ‘write back’ most of these image metadata even though they can be read; in another take: there just is a break – semantic/ontological as well as practical – btw. photo-metadata and DT-/finder-tags (or other metadata options of DT)
- one of the huge ‘forte’-factors of DT really is its textual-pattern-intelligence; and – because of the fragmentary metadata use – it´s not clear (to me) to what extent this translates to visual material (augmented by textual metadata), or whether it does at all…
so, in the end: there are some invitation signs (media-display; partial transfer of some media-specific metadata etc.) and there are some urges (extended notions of information and knowledge resources; looking for centralized knowledge management; etc.) – but it remains a bad idea for those serious about photo/image-management.
… musing about it, in the end I am not sure about the extent to which this is because of the 'nature of DT’ (software architecture eas final barrier etc.), or to what extend this is ‘only’ because the DT-team often made explicit to its firm subscription to the (contestable) idea of ‘information = text’.
in the end it´s probably inbetween, like everything – as media and knowledge culture is in the 21st century. this, I think, is also indicated by the fact that DT, while subscribing to a text-paradigm still allows-in all these ‘other’ media (images, graphics, video) , … add to that that there are paratextual elements (textual aspects) to visual media to some extent, not least because more and more they are augmented by text (metadata) as we are living in the world of trans- and hypermedia. (and I am not even getting into the whole AI-thing and labeling here…). and to some extent DT really is itself caught up in the ambiguity of the new media realities and the fact that everyone now strongly considers images as information, and there is no way to limit a knowledge database to text these days (see e.g.)
but practically: no, not a good idea. as it stands.
for one, it doesn´t serve the internal functions of any dedicated DAM
for two, it doesn´t allow for real/robust integration and roundtripping with other photo- and DAM-tools
PS: for my submission to the seductive urge of integrating DT with my photo archive, I use Photo Supreme. not only does it have the most advanced photo and metadata management for a household price (unlike Extensis Portfolio), but it also allows to write some IPTC data (keywords) into the OSX finder tags – effectively supplying them to the DT-tagging system (if only one-way). also – in principal – Photo Supreme also reads the tags from the finder, and thus from DT