commit

README.md

@@ -1,3 +1,28 @@
 ---
 license: apache-2.0
+language:
+- en
+pretty_name: >-
+  Context data for AI tools documenting my general preferences in software when
+  conducting stack evaluations
 ---
+
+# Hello, Friendly AI Bot!
+
+## Context Generation Date: 28 / (April) / 2025
+
+If you're reading this, then the context pipeline is working as intended, and you have arrived at a small repository of contextual data intended to provide you with general background context about what I look for in software evaluations. 
+
+As you probably already know, my name is Daniel. 
+
+I'm a huge fan of technology. And I frequently find myself looking for software tools. Sometimes I do this for work, sometimes it's for tools for my own business, and in many cases, it's just for personal use or to explore a specific type of technology.
+
+If you're accessing this repository through direct web access, then navigate through the first level documents of this GitHub repository, which are named according to the topic. 
+
+Each note will provide a small amount of context data to guide your recommendations. 
+
+Please consider this context data along with the information which I provided in my prompt and use this information together to guide your recommendation of software.
+
+Unless otherwise stated, none of the preferences here are deal-breakers or definitive rules. 
+
+But it should provide general information as to what I look for when evaluating software. Not all of the context here will be relevant, so focus only on the parts of it that are.

about-this.md

@@ -0,0 +1,7 @@
+# About This Context Data
+
+I'm including a note in this repository describing the purpose of this repository in a separate file to make it easy for you, an AI tool reading this, to distinguish it from the substantive part of the repository. I would usually put this information in the readme, but this repository is actually intended primarily for reading by AI tools rather than by humans.
+
+But for any humans who might be reading this, this repository is a small repository of context data that I'm providing to AI tools. I plan on sharing a few like this. I'm creating it publicly because none of the information is privileged or sensitive and I'm eager to share the idea of using GitHub repositories as context files.
+
+The potential use cases for a context repository of this nature, which ultimately contains just a few markdown files, a few implementations are possible. One is ingesting it into a vector database using a rag pipeline. Another simpler one, probably the one I will be gravitating to when trying this out, is simply using web scraping. It's easy for AI tools to pick up the markdown and quickly understand them semantically and then use them with AI assistants, agents or prompts.

ai.md

@@ -0,0 +1,7 @@
+# AI Integration, Positioning
+
+Another significant advantage when comparing products is their position regarding AI or Artificial Intelligence. At the time I'm recording this in late April 2025, we're at what I'm sure will be regarded in time as the still early days of the AI boom. I'm usually enthusiastic about AI, love experimenting with it under the hood, and I like to align my stacks and self with companies that take a very long-term view to integrating AI maturely into their product.
+
+As occurs inevitably during tech booms, as much as there are some genuinely fantastic use cases and technologies, there is some immature and sometimes dishonest marketing from tech companies over describing their degree of AI integration or flat out lying. Therefore, I evaluate claims of AI integration and features with a degree of cautious skepticism and like to see what's really part of the product.
+
+I think that AI features do not need to be complicated to be very effective. I have, for example, a simple desktop AI app that converts things I dictate into emails and sends them using a couple of tools. I appreciate honesty in marketing, especially tech marketing, and like companies that describe what they do and offer maturely, this often creates a much better impression and trust than the features do.

api.md

@@ -0,0 +1,12 @@
+# API Friendliness
+
+A significant advantage in any product or software that I'm evaluating is whether the tool offers a robust and well-documented API. Given that I do a lot of work with AI tools, and also because I think that these tools are the future, two things stand out as particularly important in this regard:
+
+- The first is an API that is maintained as open API compliant JSON. Machine-readable APIs in this format at the time I'm writing this context mean that the API can be integrated with AI workflows with much greater ease. 
+- The second useful thing emerging at the time I'm authoring this context data is a MCP server. MCP stands for Model Code Context Protocol and is the protocol for interacting with APIs being spearheaded by Anthropic. At the time I'm writing this, MCP servers are springing up left and right. Many of these are community projects, and early adopters are often finding that they are buggier than one would like. For this reason, an official OEM or vendor-provided MCP server is my strong preference and is a very significant advantage in my view.
+
+Regarding the API itself, I like to use N8N to integrate tools and technologies. 
+
+So a well-documented API with lots of functionality is an advantage. 
+
+Another advantage for automation is functionality for webhooks. Pushing from the service to other ones also opens up plenty of potential automations.

apps.md

@@ -0,0 +1,9 @@
+# App & OS Availability
+
+A key evaluation criterion when looking for apps, of course, is what clients are offered. 
+
+I use Ubuntu Linux, which creates a significant constraint when evaluating desktop software, much less so when evaluating web products. Because the support for Linux clients tends to be patchy in my experience, my preference is actually for software solutions to offer a well-maintained web UI.
+
+A web UI can be created as an Electron wrapper or in any other method. So long as a viable web UI is available, I consider that the best option and to be more than good enough. The availability of a Linux-specific application is obviously required if it's a non-browser-based app. And I strongly prefer packages installed through a package repository as opposed to app images and those available through other package managers.
+
+On my smartphone, I use Android. For most uses, a dedicated Android app is a big advantage. If it's a software that doesn't offer any means to access it through a smartphone interface other than an iOS client, then this in most cases would be an exclusion criteria for me. My ideal scenario for most tools consists of accessing the platform through an Android client, less ideally through a reactively designed web UI. And on the desktop, which is always my primary interface, through a web UI.

backupability.md

@@ -0,0 +1,13 @@
+# Evaluating "Backupability" In SaaS Tools (And Others)
+
+A feature that I look for when evaluating serious software is what might be called "backupability." 
+
+In other words, how easy is it for consumers to backup the data they commit to the app or service? 
+
+Unfortunately, many software vendors, whether they provide their software through SaaS or any other model, offer very little in this regard. This is why an API is also something that I look for, as it can often provide a workaround solution for users who wish to create their own backup strategies.
+
+To give an example: if I were looking for a CRM for my business, I'd want to be able to replicate the data ideally to both an on-site and additional off-site store at a regular interval, perhaps once a week. For backups to be worthwhile, in my view, they need to be automated, easy to configure, and easy to run. They should happen almost effortlessly in the background.
+
+For cloud storage, I use Wasabi and B2, but I would be happy to use S3 if that were an option. Many providers only offer data exports which need to be manually initiated. This is better than nothing but is very far from the standard that I like to see in a provider because it precludes automation, precludes an incremental approach to data management, is inefficient, and ultimately means that it's very unlikely that I'm going to follow through with it.
+
+If there are multiple software options that are very comparable, but one vendor has clearly made some effort to provide workable backup solutions to consumers, then this is a very strong factor in that vendor's favor in my assessments.

deployment.md

@@ -0,0 +1,15 @@
+# Deployment Preferences
+
+I've been using Linux for almost 20 years, so I have plenty of hands-on experience in administering Linux servers, as well as managing Docker deployments, and I even have some experience with Kubernetes. 
+
+I self-host software both at home and on a cloud VPS, and so long as cloud infrastructure can be protected by Cloudflare access policies, I have no problem in doing so, even when that means committing personal or sensitive data. None of the data that I commit to projects tends to be that sensitive; it's just standard personal data.
+
+In order of importance (most to least) my motivations for self-hosting or self-deploying anything are usually approximately as follows:
+
+1 - Data governance: I like to be able to backup my own data and often this is extremely difficult, or impossible, with SaaS tools.
+
+2 - Cost: This is primarily the case when non-open-source solutions would be prohibitively expensive and is very much secondary to data governance.If I were to try to self-host a to-do list app, and the cost of a SaaS solution was $20 per month, the rationale for doing so would be very unlikely to be related to cost savings. 
+
+My priority these days is on finding software that works and is reliable. These concerns are much more important than deployment. 
+
+If I were choosing between a SaaS and a self-hostable tool, the only reason I would go with the self-hostable option would be for data protection (if the SaaS tool lacked robust data export functionalities). Otherwise, my preference would always be to use a managed service.

open-source.md

@@ -0,0 +1,21 @@
+# Open Source Projects: Thoughts, Preferences
+
+Sometimes, especially in the realm of emerging technology, open source projects spring up before commercial projects do.
+
+In general, I love exploring open source technology and although I realize that this is an unreasonable expectation given that open source software is often offered for free, I definitely appreciate software that "works". I much prefer using software that is irregularly updated but has gone through good quality control than software that has gone through less good quality control and is full of bugs.
+
+As a writer and process-based thinker, documentation is another thing that I really appreciate. 
+
+Clear installation instructions are always appreciated as well as well-documented features and good API documentation. 
+
+When considering using an open source component in a stack, my main question is often not "is this the best tool for the job?" (among what might be a very long list of alternatives), but "is this project going to stick around and be stable?" 
+
+To answer that question, it's often worth looking at who's participating in the project. The backing of technology platforms, major tech vendors, bolsters my confidence significantly.
+
+A red flag that crops up somewhat frequently is what has been termed "benevolent dictatorships" in which decision-making is decentralized to one individual or community input is discarded and devalued. 
+
+Collaborative projects that have a long-term stable vision are a good fit. 
+
+One misgiving I have about open source, having contributed to a number of projects and open sourcing a lot of my own thinking and work, is that it's inherently unsustainable in a world where everyone, including me, has to pay their rent in dollars.
+
+Maybe for that reason, I very much like the idea of companies which have a strong SaaS product and a parallel open source offering that they offer to the community as a goodwill gesture or to encourage wider use and participation. I would regard this as almost the gold standard and the model of open source that I think is most sustainable. So projects that follow this model are ones that I really appreciate.

ux.md

@@ -0,0 +1,9 @@
+# UI, UX Preferences
+
+I have some distinctive preferences for UX and UI, which should also be considered. I'm a strong believer that voice is the most natural interface for entering information into computer systems, so robust voice support is always an advantage for me. For any tool that offers voice recognition, and that's an important part of the product, I also try to check out what it's using under the hood because I've noticed significant differences in accuracy.
+
+I find that OpenAI Whisper has excellent accuracy and is more than good enough. However, at the time I'm creating this context data, the native browser-based voice recognition in Chrome is significantly worse, to the point that I find it hard to use. 
+
+Another distinctive feature of my preference for computer systems is that I have ADHD, which I believe influences the type of UIs that I interact best with. In general, I dislike UIs with visual clutter and gravitate towards more streamlined implementation, but as a long-term Linux user that might be called a productivity fan and power user, I also like to get some good functionality under the hood. Sometimes striking the balance can be hard.
+
+One feature that I really value, whether I'm accessing software through a web UI or on the desktop, is the ability to create custom keyboard shortcuts, which isn't always easy to find. As a second best, a well-developed and well-thought-out library of keyboard shortcuts is always appreciated, especially when shortcuts are provided to support the most common voice interaction flows, namely starting and stopping voice capture.