Dear All,
As a recently joined member and having added a few files and "bitprints", I wanted to take a few moments to reflect on what I've learned. I must firstly praise the worthy concept behind this site. Inevitably with a note of critique there is a tend to focus on the negative or the deficient and in some ways that isn't fair. The concept and the potential for this site and for the community that it will increasingly serve is enormous and any comments I make here should be seen in the context of an approval and desire to improve and not as stone casting at a poor effort.
This could become the Google of digital media or the so-called "transient web". This is extremely important and the Bitzi developers and the user community can shape that into being. I'm sure Bitzi would be an excellent investment if the liability could be limited and if such investment was being sought. In that light then, I have a few comments.
The web pages are really rather "busy" visually. Lots of advertisements, lots of verbiage. The visual of a ticket is quite cute but the comparison with Google is a valid one. A search page resolutely bland with just a couple of ads at the top or down the right hand side with limited contrast and two or three line search responses in each case. In the context of Bitzi we might then have links that lead to the ticket page but with rather less visual "action" therein. Everybody respects the need for Bitzi to generate advertising revenue but the present layout seems to detract from the usability and that will eventually defeat the goal (not least of the advertiser).
The site itself is very slow. This has been noted before in the forums but I have seemingly lost bitprints to responses like "page timed out" and this does need to be fixed if the level of traffic anticipated is to be maintained. In fact it has taken me hours to submit this message this evening.
I see that rated filenames are promoted over unrated ones even when these arrived first. Perhaps to encourage people to stand by their submissions, anonymous contributors should not be acknowledged as the first contributor on the search page once a non-anonymous one commits?
I have some suggestions for further tags:
1. When someone adds content which they themselves author, they need to be able to notify if that content is then revised. Typically a content item like a document is added, then subsequently a new item is added, and the old version is given a "revised by" tag pointing at the new version, and the new version is given a "revises" tag pointing at the old one.
Clearly we want pristine files on the network and these tags are not meant to encourage people to submit content that is half-baked. But in the real world sometimes revisions are unavoidable and this provides a simple but convenient means to relate files so that people downloading can be assured they are getting the most recent version.
2. Alternatively there could be a means by which arbitrary files can be related along with a reason why. I note that a good attempt at automatically deriving some file relationships is already done on Bitzi. For example by noting when more than one file has the same length and first 20 bytes. I suppose it might be possible to use the related URL feature by pointing at another Bitzi URL. I'm guessing the latter though is designed to enable users to relate further points of interest to the media. Instead perhaps this should simply be a bitprint link, along with a reason for the relationship. Some reasons might be specific options from a drop down list such as the revisions tag above. A previous poster noted a desire to relate tracks on a single CD.
3. Objective description is rather a catch-all term and it isn't immediately clear whether the objective description for a Harry Potter trailer ought to be "Trailer for the Harry Potter Chamber of Secrets film" or "Sorensen video encoded trailer; 2 minutes 5 seconds". I can understand and appreciate the desire not to use arbitrary tags unless it proves absolutely necessary but that means we need to tie down many of the tags explicitly. Full marks for the pop-up help where it appears. This proves a boon when deciding what to include and how to rate.
In the future we want to see much more embedded metadata extracted from the file itself (and I note with satisfaction that at least in the case of audio, a separate hash is kept of the data not including the metadata "track"). With that in mind tags like "Codec(s) used", "Encoder" (the name of the person credited with making the file rather than the content), "Title" (i.e. from the embedded data, not from the filename), "Author" etc. might be very reasonable. Kazaa's most recent version has a very basic set of these for each media type.
4. So far I haven't contributed a huge number of bitprints but what I have added has been mostly films. With this perspective I can see a use for tags like "Play length", "Language" (plus languages for dubbing, subtitling and narration where appropriate), and possibly "Age classification" but I'm not sure how well that would work in practice since it may not be possible to know what rating a downloaded file has unless it's included in the file.
Certainly films have a genre collection all of their own which doesn't seem to be reflected in the current list. http://www-sul.stanford.edu/depts/ts/tsdepts/cat/docs/cat_policy/nonprint/genre.html contains what it amusingly describes as a "basic list" but from there one might cull a decent selection.
5. Documents, programs and possibly multi-part archives will also need a set of metadata. Especially in documents, we will need to maintain a multiplicity of similar metadata. It is typical in professional documents for example to have more than one author listed. We either must fix a way for these to be structured within a tag, or use more than one tag. In the latter case, it must be clearly understood that further tags are supplying more correct data, not correcting previous tags. Presumably the rating system is to be employed for this purpose?
I think the Bitcollider application so far is not at it's full potential although enough has been done now to marvel at how it will shape up in the next few iterations. Adding more hashes (Freenet certainly springs to mind), enabling more file-embedded metadata to be extracted by the tool and permitting users to type some Kazaa-like details in prior to committing are obvious improvements. And perhaps most importantly a Mac OS version, or at least a Java one. A significant amount of "legal" content production is undertaken on these and Silicon Graphics systems and having a port will enable more serious content producers to take part. For an encore there's always a Sherlock plugin...
There are very real and justified reasons why organisations would use file sharing applications instead of or as well as the web, to reduce their bandwidth (by sharing the task of downloading with other sites), to increase reliability (by reducing the single-point-of-failure characteristics of the web) etc. I'll close on the upbeat I opened with, Bitzi is possibly at the forefront of a media revolution. Excellent work and very best wishes for the future.
Now then, about that investment...
mystyc.
