This is a huge and complex subject; it has been discussed extensively and keeps being discussed. I mention a few aspects:
1. The idea of standardizing the format was good. Adobe defined an "initial standard". Unfortunately, instead of widely discussing it, Adobe threw it on the market. There are some problems with details, and they could be mended, but in the meantime people converted millions of images into this form.
2. Adobe now tries to pass the buck to ISO. There are talks about accepting it as a "true standard".
3. Adobe believes (? I don't believe, that they believe it), that the DNG specification is enough to cover everything. However, the camera development did not stop and won't stop due to a fixed standard, i.e. the standard will have to be updated time and again, which takes time.
4. Camera manufacturers are not prepared to disclose their new developments (not even to partners like Adobe under a nondisclosure-agreements). Consequently, a camera may appear on the market with some new feature, which is not covered by the current standard.
- the DNG specification still does not cover the "dust delete data" of newer Canon cameras.,
- the Highlight Tone Priority feature is unknown in DNG (the automatic adjustment of the exposure is covered by the specification, but the fact that the indicated ISO is not the actual ISO is not covered).
Consequently many raw processors need the original camera specific information, the structure of which is unpublished. Therefor DNG more of a common framework than a complete standard.
Still, the idea is good and I hope it will gain momentum, if ISO takes over the standard (and mends it).
However, the advantages are irrelevant for those, whose camera does not produce DNG. Plus, an unfortunate circumstance, even some of those DNG files created in camera are messy. As a general rule, camera manufacturers ought to refrain from touching software issues; they are suckers, particularly the Japanese.
Now some practical considerations:
1. Most raw processors support now DNG, but it does not help anything if Adobe updates the DNG converter only with new releases of ACR. Even if a sensor is identical to a former one (like Canon 30D after the 20D), the converter program code needs to be upgraded. It is plainly bad design. (ACR too needs to be updated, but that's the pain of ACR users only).
2. Some people are keeping (archiving) *only* DNG. There are some supporters of DNG, who are blind and deaf (but not mute). This is not the question of the standard, but of the actual converter program. Adobe recalled ACR 4.3 a few weeks ago, together with the associated DNG converter, because of errors in the conversion.
It is possible to keep the original raw inside the DNG. IMO this is a ridiculous way to preserve the data.
3. The fact, that the DNG compressed images are smaller than some others is conceptually irrelevant. The matter of fact is, that Adobe adopted the very same compression method, which is used in the Canon CR2 files (namely the original JPEG lossless encoding). The file size advantage comes from a rather lazy approach on Canon's side (did I mention, that they suck in software?).
4. Recording the adjustments made on the raw data is an important consideration.
Canon's DPP appends all adjustments to the raw file, or they can be saved in separate files, though the latter costs extra steps.
ACR can save the raw adjustments in sidecar files, or in the database; however, in case of DNG, the adjustments will be inserted in the file.
There is no generic rule, which way is better, for it depends on one's workflow. However, ACR's DNG handling makes it practically impossible to keep several versions of adjustments. Note, that this is not the disadvantage of DNG but of ACR.
5. Adobe's DNG converter is a stand-alone program, hopefully without a time bomb. I keep all previous versions in my archive (even though I don't archive DNG files :-). However, the program source is not open entirely.
On the other hand, should Adobe stop supporting DNG, someone would certainly jump into the niche, so there is no reason to fear for the support of old stuff.