Boswell - 4.0.1

Before I start, I should mention that I use Boswell in a manner which is different than its designers use it. I am a believer in using a plethora of tools to get a job done, where each tool is designed to do its one task best. You could probably cut a board and then nail the two pieces together with nothing but a hammer, but you'd have a pretty messy "cut" and probably a sore hand. Boswell has enough text authoring tools to get simple things done. But, I primarily use it for archival, and that is where these comments will be coming from, philosophically speaking.

I have done a lot of looking and researching for a program to do just what this one does. So I will answer some of birame_dotmac's questions. I should say that most of this is very much from my point of view. There are no absolute answers to some of these feature questions that I address. Everyone has their own needs and work styles. For me, I am comfortable in a text-only interface. Not everyone is. I write these to explain what a text-only person can do, and to hopefully convince a few people who are sitting on the fence over whether or not they really need Feature X. All of the alternatives that others have listed here are, in my own opinion, software of a different hue. Boswell is a _data_ archival and retrieval system. Let's begin:

File Types

Do I need to actually archive PDFs? No, I don't. 99% of the time, I want the data /in/ the PDF archived, and I could care less about what the typesetter wanted that data to look like. Data stuck in a PDF file is extremely limited in its uses. For that one time in a hundred when format is necessary, I can move that file to the archive folder, and create (or add to) a gloss page in Boswell with a pointer to that file. If it was a PDF I made, well, the same goes for that. I archive the LaTeX mark-up, not the end product. The same goes for photographs. I already have the file on my hard drive; I have an iView catalogue entry of that file with more meta-data than any SuperNotebook application can dream of providing, why dump the file into my archive? Why not just provide it in a static location and point to its location like any other index in a library? It does not bother me, and it keeps the format tidy and slim.

I had decided on that long ago, before even finding Boswell, mostly based on the fact that the way most applications handle embedded images greatly bloats the file size. Additionally, philosophically I am a believer in separating form from data. Structural data, not stylised data. So for me, many of the features these other applications provided were superfluous, and only got in the way, slowing down development on the areas of the application which interested me.

Links

Links, on the surface, are a missing feature. But, I differ on this -- allow me to rant a bit. Over the many years of using computers as a writing tool, I have evolved a system of internal references which works great for me, and relies on no tools other than a text editor. Back in the day, that is all we had. Links seem to solve a lot of issues, but on the other hand, they /create/ some very serious issues. I have switched primary development applications dozens of times over the decades. Software gets obsolete, goes out of business, what have you. What happens to all of those links when you are forced to drop an environment like that? In most of these applications, you lose everything. The link disappears, and no alternative reference is created! This is a very critical flaw that many people are not anticipating. What if you only export part of your database, do links with missing targets silently disappear, or do they point into nothingness? Can you /search/ for your links? Since mine are text based, I can. I can search for link type and target, isolate all matches and place them into a target area for analysis. Afterwards, I can do a global search and replace to set the link types to a different development stage. I don't know of any SuperNotebook apps that can do any of that. Removing these text links for print is a simple matter of search and replace.

I know it grates on people to be denied things that we are only getting a taste of in the modern computing world. It is easy to forget that conceptual programming should be considered Beta. It might not always work, you might /lose/ your work. While it is fun to play with applications that have whiz bang features -- these features are not by any stretch Stable. Now, for an application designed to lay out thoughts quickly and temporarily, that is fine. Take advantage of the year 2005 as much as you can. But what we are talking about here is an archival application. This is the kind of application you will be storing your documents in for decades. I would rather stick with the most conservative forms of conveying data as possible, and plain text like it or not, is going to be around for a long time.

I do believe fine grained linking will be a basic data feature of the future, a stable feature, if you will. When -any- application can bulk find, modify, delete, move, export to X format, a link -- then I'll switch my "old fashioned" links in one big set of search and replaces. Until then, I'll stick with what works in the long term.

As for links to web pages. Having gone through several years of documents already, I've noticed that the things I provided a URL for in a document, generally are gone. Linking to the web is nice for temporary documents, but for archival, it generally becomes more useful to locally mirror whatever resource you wished to reference, or to just describe it.

Multi-level Hierarchies

The better question to ask is whether or not multi-level hierarchies are an adequate system to fully organise data. The fact that most good hierarchal based application have a feature called "clones" or "aliases" is a clue to its limitations. Furthermore, most applications treat these clones as inferior to the parent node. I have seen two applications deal with clones intelligently: An outliner called TAO, and Hog Bay Notebook. In them, a clone is indistinguishable from the original node. If you delete the original, the "clone" becomes the original, from the user's point of view.

Hierarchy does have its uses. It is useful to say that documents X and Y are both like A -- but is there a sound reason to display that connection in a strict tree -- especially when Y can also be like B and sometimes C but only when coupled with Z? You see what I am getting at, I am sure. Mental data, and that is what most of these applications attempt to capture -- is far too complex and dynamic a structure to be displayed in a tree. Trying to do so eventually goes one of two ways. You either simplify form classification to the point of /long-term/ uselessness. Or, you spend all of your time classifying documents.

As I am sure you know, since you brought it up, Boswell only has three "levels" of hierarchy, in the strict sense of the term. There are notebooks, entries within notebooks, and entries that are not in any notebooks. You could also say there is a fourth, entries in the Ignore notebook, which is essentially a trash bucket.

Instead of having the user classify documents into a tree, it uses a somewhat unfamiliar, novel system which I've only seen successfully deployed in two other places, Google's Gmail and Opera's mail system. Gmail, like Boswell, simply dumps everything you own into one Archived area. Labels and searches allow you to discover these when you need them. It was what Apple /wanted/ to do with Spotlight, but they have yet to achieve that goal. Like Gmail, you can have Boswell automatically "filter" data into pre-defined labels or notebooks, as it appears. Also like Gmail, but unlike Spotlight's Smart Queries, you can manually edit the contents of these destinations.

The similarities depart, however, when you consider Boswell's Manager system. This is an extremely powerful tool, which allows you to funnel data around the archive, change meta-data, and such by using extensive Boolean queries (my only gripe there is that you cannot save these queries for later use). Gmail has no such power, yet.

So, I can make a filter which automatically detects the presence of my email address, preceded by "To: ", and sorts everything with that string into Email-Received_NB. Later on, if I discover a batch of emails sent to an old, forgotten address of mine did not get automatically added, I can use the manager to root these out of the archives and append them to this notebook as well. Simple as that. Found a bunch that got added by mistake? Remove them. This is what Spotlight got wrong.

The smart notion in this system is that an entry does not really reside within a notebook. The notebook is a lens through which you can see parts (or all) of your archive at once. When you are done with that lens, you can delete it, or sort it to the bottom of your list. Like a lens, you can see the same object from multiple vantage points. The same exact entry can be viewed from multiple notebooks, and like a lens, can be distorted via its Tag to be relevant to the focus of that lens.

Clippings

I know some people really like this, personally I never found much of a use for it. Clippings are typically collected into some central repository of such things, and thus require further sorting on a regular basis. I might as well just drag and drop, or copy and paste, directly into whichever document I wish to place the text, wherever I wish to place it -- and then forget about it. So for me, it seems to be an extra step. It usually involves extra work to clip it. If there is a keyboard shortcut provided, it is non-standard, require memorisation, and hopefully does not conflict with the source application. If it does not, it requires an awful lot of mouse usage, and gazing at a horribly unorganised dumping ground for Services.

Visual Layout

I have already expressed my views on structural documentation versus styled documentation, a bit. I am a pretty strong believer in divorcing content from layout. The primary reason is flexibility. A document, if structured properly, can then be exported to several meaningful formats, and the details of those formats can be adjusted at whim. Why set in stone that all headers must be this font face or that, when I can just say: "This is a header." and be done with it. Most writers are never faced with the issue of layout, anyway. Some style dependant programs are capable of displaying flexibility in this manner, but rarely to the extent that I feel is wise, and very few of the so called data managers are good at this.

The second reason comes from the same reason that I do not opt to use current technology linking: Portability. I do most of my writing in either Ulysses, or TexShop for final tweaking. These produce two plain text files, one with Ulysses structural mark-up, and the other with more complex LaTeX mark-up. Ulysses mark-up is light enough that it does not get in the way of inspiration at all, and Ulysses can export to LaTex. So I rarely feel the need to archive LaTex files. Since Ulysses files are plain text, Boswell handles them just fine, and if I ever want to revisit them, I can just copy and paste the source back into Ulysses for further development -- and then re-archival as a new version.

That is the most that I ever do with styling. To add advanced styling to Boswell would be something I never use. For me, it is simply a place to store portable documents "forever". It is not, and should never be, a typesetting engine. There are plenty of industrial strength applications and system already in place for such styling.

I have played with several applications in the past that allow extensive placement. Curio is one of them. While I found it to be a very intuitive application which allows inspiration to flow, I would never, ever, archive anything in that format. For me, there is an important distinction between working environments, and storage environments. I use the best applications that I can find for each purpose.

Export to Various Formats

I already touched on this above. Boswell's primary export is a simple text file dump utility, either on a single entry basis, or by entire notebooks. I would not use Boswell to create a web site, I use Tinderbox for that. Best tool for the job. I certainly would not try to create a web site with DEVONthink, either, or many of the other listed options. But as for the relevant, save-worthy content of the site? That all gets archived in Boswell.

Versions

You did not mention it specifically, but it gets brought up a lot, so I will say my bit on this, too. I would say that the inability to delete or edit entries was one of the main things that sold me on Boswell, right beneath dynamic, unforced filing. I think the thing that worries people the most, when trying the application, is this immutability. Coming from a background of digital common sense where everything is malleable and volatile, you are used to making silly mistakes because they can be easily fixed later. Few remember that with Boswell, you needn't archive something instantly. You can leave it in the Journal for up to an entire month. That is plenty of time to get all of the tweaks worked out of it, and by the time it is ready and sitting around for a while, Boswell silently archives it and you are done. If you do ever need to change something, you can make a copy and make changes to that copy. Et cetera. You are then free to forget the original, or keep it around for reference.

So, practically speaking, it would be better to think of Boswell as having forced versioning, instead of thinking of it as being "un-editable."

I don't think I have deleted important files too often. That is something I am pretty careful about, but keeping old copies of a working document? There are thousands of stores and articles I wish that I could go back and few the history of. Boswell just makes sure you always have access to those versions -- always. If you really truly do not need the version, move it to __Ignore and forget about it. It is as good as gone -- but not *really* gone, and that is the key.

Interface Awareness

The ability for an application to recede into the background is important for data and particularly mental data storage. When it comes to inspiration, you don't want anything in the way. Boswell is one of those rare applications that truly gets out of the way to the point of barely existing, until you need it to. At its most hands-off approach, after setting up a few filters and notebooks, you could write in the application for a few years and never once archive a document, or worry about where it is going. The filters and auto-archive at one month feature will work in the background, until one day you need to look up that article you wrote for such and such magazine in 2001. So you run a search for entries in the freelance notebook you set up four years ago (and forgot about) with results returned for items stored in the year 2001. One instant later, you have a neat list of articles, and the one you had forgotten about is just as prominent as if you had carefully stored it; labelled it; and cross referenced it. You find yourself looking up a lot of these articles in the next month, so you set up a notebook for freelance 2001-2003. and tell Manager to dump the results into it. There, it is as if you had been carefully indexing these things all along.

Now, most people are going to want a slightly more on-hands approach than this. If the application can provide that level of control over such a haphazard style of usage, just think of how useful it would be to someone who has spent a few minutes a week making sure everything is organised in a way which will make sense in ten years.

Believe me, the reason why my writings are so unorganised is not because I am an unorganised person. The reason is that every time I have tried to use an application to sort my writings, I have run into a brick wall of micro-management, excessive labelling and structuring, or the realisation that I cannot even locate things I stored a few weeks ago. Boswell is the first application that has allowed me to input old data with blazing speed, and the confidence that I'll be able to find everything later.

A few words on each application alternative mentioned

Notebook/Notetaker: Too limiting for me. I don't like the idea of storing massive amounts of data in what is essentially a limited depth outline (with very strict hierarchy). I have a huge quantity of written material that needs organising. There are some really nice features in these though, and everyone should check them out.

Brainforest: Same as above. An outline is too limiting. Just putting one novel in an outline would be a tedious exercise, and would tax the file format. My Boswell archive has two novels in it so far, and over four years of daily diaries. That is a fraction of what I need to import, but it loads instantaneously, saves instantaneously, and searches instantaneously. I am not concerned about how robust it will be in the end.

MacJournal: I have ethical issues with the development of that software, for one. It is a nice application, and I did use it for a while before it got sold to a company that I'd rather not have dealings with.

Stickybrain: I only tried this one briefly. It struck me as being something more of an information gadget than an industrial document storage tool; not sure what it is doing in a list of alternatives.

DEVONthink (Pro): I actually went back and forth between the beta for Pro and Boswell before eventually making my decision. Out of everything here, for archival storage it came the closest to being a contender. In the end, I had issues with its over-all design philosophy. I felt somewhat hampered by the interface. I cannot really describe it much better than that. I had no rapport with the software. It has some definite positives over Boswell, but in the end I preferred a simple elegant solution to a complex solution. I also like how Boswell only filters the way I tell it to filter. With DEVONthink, I felt highly out of control, as if the software thought it could do a better job of filing than I could.

Tinderbox: I am an *avid* user of Tinderbox. I use it for development on many different projects. My massive, fully automated To Do list; all of my notes and drafts for my current novel; my web site; the web site I maintain where I work; and other diverse sundry little things that need a big brain to sort parcels of data. On the massive document storage end of things though, it is less suited for that. Prior to revisiting my search for an archival system, I did use Tinderbox to archive my diary for two years. It is a large file. Load and save times are definitely starting to tax the limits of that file format. All of that data is now in Boswell, and as I said, it feels just as snappy as when it had nothing but the default documentation entries: A fresh database -- and the Tinderbox file did not even have the novels and other significant chunks of data that are currently in the Boswell file. So, for what Tinderbox does, it is superb. I can maintain my site with templates and logic rules, making it so that I never have to worry about links and whatnot. Pages are HTML-less notes which can be dropped anywhere in the system and adjust accordingly. It is brilliant for that kind of system, but overkill for archival -- and not really designed with it in mind.

My Ratings

Ease of Use: I see a lot of how this application is complex and has a steep learning curve. I think a lot of that might have to do with reviews of older versions, which had a lot of unfamiliar jargon. From what I can tell, this jargon is all gone. A Filter is pretty obvious for what it does. Personally, I found learning the technical aspects of Boswell, a breeze. The fundamental philosophy is a little more difficult, but more because we are all so used to living in a volatile digital world, not because the application is extremely complex. I gave it four stars out of five. I think some events could be slightly more intuitive, such as drag and drop, but over all, when compared to things like Tinderbox, FileMaker, or even simpler applications like MacJournal -- I think the complexity of Boswell is more one of those things that has become a self-perpetuating given, based on old reviews. I could just have had an unusually easy experience, however. A meshing of ideologies.

Support and Documentation: Support has been fantastic in taking the time to write well thought out responses to my initial questions and concerns -- even before I was a paying customer (which was not for very long! :) As far as documentation goes, I think it speaks highly of an application if its documentation can be readily inserted right into what the application does. Having those articles right in your library is very handy. The writing style is light, comic in places, but informative. It spends a lot of time on philosophy, which I think it good. As I said above, I think that philosophy is going to be the sticky point for most people. Five stars.

Features: Five stars for a generally solid application. There are some additional things that I feel would make it even better, as well as some refinements to what already exists, but in its current state it is very powerful, and even if it stopped moving today, I would have no regrets in making the purchase I did. I cannot say that about many applications.

Quality/Stability: Four out of five. Really four point five. While I have had a few crashes, there has been no data loss or corruption. With something like this, that is what is most important. Quality is high, though. This is not something that feels thrown together, or as if it has barely supported features. Every feature is very solid, and put there for a good reason.

Price: It's a little on the expensive side, I will grant that. I've found that the types of developers who charge more for their product seem to take stability and conservation of features a lot more seriously. Cheaper apps tend to have a lot of features, but they often lack a sound basis, good UI, and are often buggy. I know, it is a generalisation, but of the three main writing tools I use, Tinderbox, Ulysses, and Boswell, they all cost well within this price range, and they are all very solid, well thought out applications with careful developers who are passionate about what their product can do. It takes all of those factors, with skill and talent of course, to make a good application. I am willing to fork over the extra $25 - $60 (based on the alternatives listed) for that kind of product. Five out of five, not because it could not be cheaper, but because it could be twice the price and still be a pretty good deal.