When I first switched to the Mac in July, one of the first programs I started using was DevonAgent (DA) which I viewed as a replacement for CopernicAgent (CA) which runs under Windows and is the only program that I know of which is comparable to DA. At that time, I felt there were a number of shortcomings in DA compared to CA which I posted about. After that post, Bill advised me not to compare DA to CA without taking into account the capabilities of DevonThink. After using both DA and DT for some time now, I see the wisdom of his advice, realizing that DA and DT are so tightly integrated that I consider them to be part of a single package. So, instead of trying to make a comparison of DevonAgent to CopernicAgent, I thought I would take another stab at a comment on the whole Devon package which I will just refer to as “Devon.” I am posting this to the DA section because most of my suggestions involve DA.
Before I do that, I need to explain a bit about how I use Devon. My work commonly involves gathering a large amount of mostly online information on a given subject, usually an organization or an individual, and then using that information to write a report. I would estimate that about 75% of this material is from webpages, another 20% from Lexis/Nexis and perhaps 5% from paper documents which I scan into pdf files. I usually begin with a DA search using a search set consisting of the largest search engines and then dumping the results into DT. I then supplement this material by importing any relevant material from my hard-drive (which, incidentally, I search using a superb program called FoxTrot instead of the almost useless Spotlight). I then get to work using DT’s autogroup and classification functions to start categorizing the material into an outline which later serves as the basis for my report. (Since almost all my material is online, I don’t have the problems with sourcing that some users seem to have since the URL, title, etc are preserved in the DT data base.) All things considered, the Devon software is probably just about perfectly matched to the way I work and far superior in most ways to the combination of CopernicAgent and NetSnippets that I used to run under Windows.
That said, there are a few areas which need improvement:
- PDF’s and other file types
Since filetypes are than html are basically invisible to DA, they don’t make it into my initial sweep. I can’t just ignore these files for obvious reasons, so I have worked out a way to get them into DT. I do a search on Google and restrict the filetype to PDF for example. Lets say I get 3 pages with 100 results on each page. I then use the three URL’s as input into the Download Manager which then imports the files into DT where I dump them into the files with all the HTML files for grouping and classification. The lack of pre-filtering is not a problem with these kinds of files since unlike web-pages, documents don’t change and if they are unavailable, Download Manager simply doesn’t find them. This system works fine but it is a bit cumbersome. I think it would be better to have an option in DA to bring other filetypes into the results. As I said, filtering is probably unnecessary so it shouldn’t be so hard to do.
- The Funnel v. The Net
I see Devon as a kind of funnel into which you pour a lot of material and getting useful results out of the end. For example, I just started working on a project which will involve processing about 1500 documents which I can’t imagine doing without the Devon capabilities. However, there is situation which I sometimes face that is the opposite of dealing with a large amount of information. Sometimes there may only be a handful of online sources on a given subject. Hypothetically, lets say that I am researching Mr. X and there are 7 sources on the entire Internet broken down as follows:
Six html documents, two of which are only in the Google cache
Three PDF documents
One .doc
DA is only going to find 4 html documents and will miss the cached HTML and the PDF/doc files. Once again, I am going to have to run a separate Google search to make sure I don’t miss the other file/types as well as the cached files. For some people, missing a few documents might be irrelevant but for me, in some situations, it could be a catastrophe. CopernicAgent did find the cached files (as well as the other file types) because the filtering unction could be turned off, instructing the software to return all results, even if the pages did not contain the search terms. Again, the work around is to simply run a Google search in these instances. By definition, there won’t be many files so importing into DT is not a big deal.
- Evaluating Plugins
I don’t see any means to identify which plugins are returning which results in DA, so how do I know which plugins are most effect for my purposes? If there is a way to do this, please let me know but as it stands, I find myself just guessing about the efficacy of the various plugins.
- Archives
For me, the archives are almost useless. My searches are often highly related and I usually will want to see a document even if it is in another archive because that document will be relevant to current research. The only time I want to avoid a duplicate is when I am updating a particular search. So, best would be the ability to filter against a particular archive/s. I could work around this by searching the archive first, but it is an additional bit of work and as it stands, there are no boolean searches possible for the archive. There is another workaround involving making backup copy of the archive and then deleting all but the relevant archive but that is both cumbersome and potentially dangerous, risking losing the whole archive.
- Stability
I find both DA and DT to be a bit quirky. DA almost never actually finishes it searches for me and DT seems prone to corruption in the database as being subject to the eternal spinning beach ball for reasons I don’t understand. None of this crippling but it can be annoying.
- Manual
I consider myself fairly intelligent, but often I can’t understand what the help manual is saying. For example, I just read the section on workflows and I still don’t understand what most of them do. This thing needs to be rewritten.
So, all in all, I am very satisfied with the switch now that I understand how DA and DT work together. I would like to see the suggestions I made incorporated but I can basically live with things as they are now with the exception of the archive problem. This really needs to be fixed, ideally with the selection option or at least the possibility of boolean searches. Also nice would be the addition of some kind of visual metaphor for organizing in DT, but thats a big subject and probably would need another post to discuss.