Log in
Memory Health: Alzheimer's Support Group
836 members365 posts

Project: WikiJeff

Hi gang, Jeff Cobb here. As my disease progresses, aside from the usual fun stuff with walking and memory, I find that those windows of time where I can more or less focus only a few steps below what I used to, these windows are coming less and less frequently and at pretty random times. I am finding that times that I can focus enough to even write in this blog only come once in what feels like a while. Its not that I don't want to write in that state, I simply cannot figure out how to operate the pretty brain-dead WordPress dashboard. Where before my idiot-cycles or low-phases consumed maybe 10-20% of my waking day, now its more like 50-70% of the time. In this state I can barely operate the apps I need, if I can even turn the thing on and log in properly at all. Memory loss sucks because this is my one primary remaining contact with the outside world and even if memory loss prevents me from operating my box correctly enough to get done what I want, its never strong enough to wipe away the shame and sometimes guilt at not being able to accomplish a simple task or seemingly falling out of touch with the very few people I still interact/speak with.

One of the side effects of this is that I cannot predict when I will be up to doing any given task, and more specifically writing. I feel the need to write down as much stuff about myself as I can so that when I can no longer make sense when I speak, I will have a record or document that can speak for me. This isn't a someday thing either; I know for the fact I will be needing this or something very much like this soon.

So far I have simply kept a simple directory of text files I write on different subjects ranging from my youth to how to make hashish to speculative AI software design. One day I might feel I can write *something* but by the end of the first cup of coffee I know that "something" definitely won't/can't involve techie stuff for instance, but I do feel like I can jot down background stuff about me, which is more in line with simple documentation of events whereas pretty much everything else requires most of my brain being functional to work.

The drawback to that is there is no organization, no way of finding any specific line or bit of information, some things have close duplicates, etc. So after noodling things over and taking my unpredictable cognitive functioning into account, I have decided to make a wiki of myself. This way if one day I feel that writing about technical things that help keep our house running I can go directly to that section and start adding information. On the other hand if I feel the best I can do is write down my special chili recipe, the wiki is ready for that as well. Where plain documentation is very linear, the wiki model is extremely free form.

With that in mind I set out some requirements for any software I might need to use:

* Has to be stand-alone, no external web server needed. This requirement actually satisfies a few things; the obvious of course but also, a standalone system of any kind is inherently easier to operate...but just in case...

* Has to be simple enough to operate in any state I might be in. In fact it must be able to start up and shut down with a single command.

* Has to be free, both as in freedom (open source) and as in beer. I have no money to spend on things.

* Has to exist within the universe of my Linux package library. Allow me so explain: all software on my flavor of Linux uses a packaging system for downloading, installing, uninstalling, etc software. Right now in this universe there are some 25-30,000 packages or programs available for download for free. The good part of getting your software this way is that it is so easy to use and damned-near foolproof.

* Has to be so easy to use I can call one command to start it and then just point my web browser at it and start interacting with my wiki.

I found an open source package called "didiwiki" that seems to fit the bill. It is light weight, comes with its own server and I was up and running in minutes. Installing software on Linux is vastly different than it is on Windows or Mac. For example, when I wanted to see all the wiki apps available to me I just had to type in a terminal:

>apt-cache search wiki

And several screens worth of wiki applications and tools streamed by. Had to narrow down since my LBD will put me in brainlock if I stare at a list like that for too long. So the new command:

>apt-cache search wiki | grep simple

Note: the above command says to search through all packages available to me with the word "wiki" in the name or description, then the grep command says of those returned, only show packages with "simple" in them as well. This time the list was manageable and I quickly settled on one called "didiwiki". Didi means quick in Vietnamese I think. Once selected, all I had to do in order to install Didiwiki on my system so I could start using it is enter this:

> sudo apt-get install didiwiki

What that means: "sudo" lets me borrow admin privs for a single command, apt-get is the main installer app, we tell it we are doing an "install" and the name of the package we wanted "didiwiki". Thats all it takes; the system checks things on my laptop, then goes out into the internet and finds the best version of not only Didiwiki but all support apps and files as well, downloads them to my laptop and installs it in my system. Seconds later it is done, so being a good UNIX dude I checked the man or documentation page for how to start it:

>man didiwiki

This returned all the many ways you could use it but the most important part of those docs was something that went like "if all you want to do is get this going, type this:"

>didiwiki -l

That starts the wiki software listening on the network address of which is the default "home" address on any computer, it always means "this machine." Once that starts the app, I check it with my web browser to see if its working by entering the following in the address bar of Firefox:

That means look at this machine and on port 8080 for a webserver. 8080 is just the default that comes with Didiwiki. I do that and presto!

(see picture at top of story)

Ta-dah! My very own personal wiki that I can now populate with the detritus of my mind..

At first this will be pretty damned useless and may remain so for a while as information builds up within it. However once completed, temporary caregivers will have a simple way to check my needs etc right on their smartphone; the search feature alone will help not just me but my caregiver and doctors find needed information in a timely manner. For example if someone new is watching me and my day is "bad", she could ask me about allergies for example and I might stammer and never really get a good answer out. Conversely, if she points her smartphone browser at (actual address of the machine) and then enter "allergies" in the search bar; instantly she will have a concise and clear list that she can be sure is complete.

I am doing this now because I can feel the days where I can do this sort of thing drawing to a close, and I have info she will need in time that I know isn't written down anywhere....




5 Replies

Great idea, Jeff. A searchable personal wiki!


Cheers Christian; actually as a fellow geek, follow me on this for a second. OK you know the problems and trials and tribulations of folks like me when we are out of just trying to deal sometimes. Its like a dead-end loop where the memory loss triggers confusion, which in turn triggers memory loss, all adding up to the dreaded "fog". So much of that can be triggered or started by not being able to retrieve key information at the right moment; I suspect its the executive functioning failing when the patient is in sensory overload combined with the patient going into mental overdrive in an effort to retrieve data like med lists or worse, med history (extremely painful for me). But ANYTHING can go into this wiki I propose because as you say, its personal. So everything about me can be put into this thing but THEN: port that to some wearable device and combine it with the equivalent of Google Glass (functionally). The Glass work-alike is keyed into the wiki via STT and responses can even be TTS. At this point if I am needing to answer questions I can because every answer Jeff is likely to need (and need to get right the first time) is stored this way and easily retrievable in public circumstances. My aphasia is bad today so you will have to let you mind wander with the possibilities here: facial recognition of family and friends as the dementia gets bad can help keep loved ones names a lookup away, my to-do list is a glance away, etc. At this point though, I gotta keep my hopes simple so at best I am attempting to:

1. Get the wiki populated on a standard PC.

2. Port the data to some Android-friendly format so its on my phone, tablet, etc. Syncing capabilities a plus.

3. Investigate integrating Google voice tools as an interface.

I think that is about all I really have brain-cells/time for at this point, at the last thing is a stretch goal to be sure.

Still, if it even half-way works it will really improve my standard of living. Right now I have Didiwiki running as a task on my house server so I can interact with any device under this roof. Seems pretty stable and (so far) does what I need.

Say question if you know: is there any kind of lowest common denominator for wikis like CSV files used to be the common go-between back when dinosaurs ruled the earth.....

1 like

Yes, .csv files can be used for export/import wiki data. Start simple, imo.


Well, depending on the storage under the hood (since this is a very self-contained wiki+server I need to dig to confirm) but using SQL to export and import data might be an option. In truth the bigger snag to my BIg Master Plan is that other tools I have written (slowly and painfully) that complete the project are all in Python and I have no experience with Python on Android. Maybe not needed if I abandon some of the more exotic voice tricks....


Hey, just had a bit of a flash: does Google publish any kind of API info about Google Home (the Amazon Alexa work-alike)? Just had the idea that if ...the hardware was hackable....and the APIs were known, the device itself could act as a front-end to the wiki, populating and updating in real-time. If such a device were possible, installing them in the homes of patients as early in the MCI as possible could (with AI and work) effectively create some type of ...record? Not simulation, although it would know alot about the patient....it would/could almost be (after a few years in MCI) not just a way to help the patient now (its five oclock, take your meds, its 09:00, your daughter is coming to visit) but it would also be a truer reflection of how the patient used to be. Words cannot describe how...important that is to someone with a progressive condition. It would also be the best description of your personality, it would have your history, your likes, dislikes and so on. And it could be of immense help for temp caregivers....