For Christmas, we bought a new iPad and our first Apple Pencil. I asked my brother, a professional illustrator, to recommend a drawing program and he suggested Procreate. I'm so glad he did. I've been having a blast. Playing around with Procreate is convincing me that you don't need a full desktop app to make great artwork. This is a bit of a paradigm shift for me. I'm still very interested in learning desktop apps like Inkscape better (check out my drawing of my dog), but it's a lot of fun to sketch and play and I love the portability of an iPad.

A while back, when I started using git, I thought it would be a good idea to keep my binary files outside of git. I created a directory called “incfiles” on my server at the time. It was for files I want to include elsewhere. These days that directory lives at http://thedurbins.com/static/incfiles. I just found a pic of some cacti in there: I was just pleasantly surprised that Hugo allows me to use non-local files as the featured image, like this:

At the recent OFA conference I met someone who has brought her copy of Coding Freedom to Harvard to get it signed by the author, Gabriella Coleman, who is a professor in the department of Anthropology. I've read a lot of history of open source but never from the perspective of an anthropologist, never as part of a serious academic exercise. I just finished the book and I'd recommend it. It's a bit academic and dry at times, but maybe it's the sort of book I should have on my bookshelf to review the dense parts now and then.

The other day I was minding my own business when an email hit my inbox: “OpenForum Europe is organising the OpenForum Academy Symposium in Boston on 13-14 November at Digital Data Design Institute at Harvard Business School. It is an academic conference that covers questions related to the social, political and economic impact of Open Source. We found your website Boston Open Dev for the open source meetups and we thought of sharing the invitation in case you and your networks might be interested.

Listening to the JS Party episode about Jerod Santo's tool box motivated me to finally get Ollama working properly. I had already played around with Ollama when I gave my friend Slava's datachat a whirl. In that case, per the README, I spun up a Docker image and ran the following to pull the 4-5 GB model in: docker exec -it ollama /bin/bash ollama pull llama3 Then Slava's app, which is pretty neat by the way, worked just fine.

Today I made a quick sketch of my dog, Midnight. He's a Shorkie, half Shih Tzu, half Yorkshire Terrier. I was inspired by all the Inkscape content during the 20th anniversary celebration today. I joined session one early this morning and then caught up on videos by Martin Owens. In one, he explained a new template he created for making zines. Flip open my zine and you see my doggo.

I watched a video on YouTube called Large Language Models and the End of Programming by Matt Welsh, hosted by CS50. I considered attending the talk in person but didn't make time for it. I'm glad it was recorded. It's really good! I first heard about Matt when he blogged about leaving Harvard back in 2010. It was nice to hear he's doing well and right in the middle of the AI revolution that's suddenly surrounding us all.

The best paragraph in Social Translucence: An Approach to Designing Systems that Support Social Processes is the last: From our perspective, the digital world appears to be populated by technologies that impose walls between people, rather than by technologies that create windows between them. We suggest that understanding how to design digital systems so that they mesh with human behavior at the individual and collective levels is of immense importance.

The first two open source meetups in Brookline were a ton of fun. We had ten and then eleven people. Due to travel we're skipping February, but we're on for March: When: Sunday, March 26, 2023 at 5pm (Google Calendar) Where: Hamilton Restaurant & Bar (in the back room), 1366 Beacon Street Brookline, MA 02446 (Google Maps) Come one, come all. If you'd like to RSVP (which certainly isn't required!

I've been in the market for a new laptop for a long, long time. I bought my last laptop in 2011 and have run Linux on it, but I've been very disappointed with the hardware (a ThinkPad X220T). I had dreams of using the Wacom tablet built into the screen to make artwork, but it just didn't work out, and it wasn't cheap either. It was time for a change. Maybe every other laptop can be for Linux.

Over the past few days, I binge listened to (and sometimes watched) all the videos from the Creative Freedom Summit that took place last week. I heard about it on the Sustain Open Source podcast. The agenda was packed with interesting talks, but here are some of my favorites: Automation in Inkscape by Máirín Duffy Inkscape's Multipage Feature by Martin Owens Creating Fedora Wallpapers with Krita by Madeline Peck Blender: Grease Pencil Speedpaint by Paul Caggegi Freelancing with Free Software by Ryan Gorley A quick painting demo using Krita by David Revoy Brand & Badges: How Fedora Designers Create with Inkscape by Marie Nordin and Jess Chitas (I'll update the links above as edited videos are available.

The first open source meetup in Brookline was a blast. Holmes and I were joined by eight other open source and decentralized web enthusiasts over drinks and dinner. We decided we'll do it again at the same place: When: Sunday, January 22nd, 2023 at 5pm (Google Calendar) Where: Hamilton Restaurant & Bar (in the back room), 1366 Beacon Street Brookline, MA 02446 (Google Maps) All are welcome, of course.

Today I listened to Syntax 548, titled “Rendering Methods Explained”, which covered a question in the State of JS survey that asks, “Which of the following architecture and rendering patterns have you used in the last year?” The survey is ongoing but an issue and a localization file show the following options: Single Page Application (SPA): Apps that run entirely in the browser Multi-Page Application (MPA): Apps that run entirely on the server, with minimal client-side dynamic behavior Static Site Generation (SSG): Pre-rendered static content, with or without a client-side dynamic element Server-Side Rendering (SSR): Dynamically rendering HTML content on the server before rehydrating it on the client Partial Hydration: Only hydrating some of your components on the client (e.

I'm co-organizing an open source meetup in Brookline about a week from now: When: Sunday, December 18th, 2022 at 5pm (Google Calendar) Where: Hamilton Restaurant & Bar, 1366 Beacon Street Brookline, MA 02446 (Google Maps) Who's coming? Well, the two co-organizers for sure. If you'd like to be added to the list below, please email me at philipdurbin@gmail.com. Philip Durbin (co-organizer), developer for Dataverse Holmes Wilson (co-organizer), creator of Quiet … This will be a very casual event.

In yesterday's post I mentioned that I was having trouble remembering the specifics of a talk from years ago. Somewhat ironically, this was in the context of a conversation about chat apps where we touched on the usefulness (or not) of old chat logs. I was asked, “When do you find yourself going back to old chat logs?” To find those specifics, I found myself going back to old chat logs!

Today I met with the creator of Quiet, a new chat app that doesn't require a server. That is, it's peer-to-peer and completely decentralized. We quickly got into how much we appreciate how the Internet itself is decentralized and both want to keep it that way. I started talking about Scott Bradner and his use of the word “anarchy” as a positive attribute of the Internet. I struggled to explain what Scott was saying but later I dug up his article, The importance of restrained anarchy.

We bought an erg, mostly for our daughter, but it took me a few days to get on it because I wasn't sure if I'd set a new personal record. Here's the data: date piece time split stroke rate 2022-07-12 2000m 9:33 2:23 26 2022-07-31 2000m 9:14 2:19 27 2022-08-06 2000m 8:57 2:14 26 2022-08-22 2000m 8:42 2:11 26 I had managed to shave off a few seconds each time. Could I keep this up?

Today I was messing around with GitHub Actions for OpenDP. I found an answer on Stack Overflow that got me unblocked. I ended up splitting up my single step into multiple steps so I could indicate that I'm ok with one of the steps failing with continue-on-error. I'm reminded that I should check out the GitHub Actions advent calendar Ed Thomson made. I got a chance to meet him last year at FOSDEM.

As of today I'm working 60% on Dataverse and 40% on OpenDP. My role for Dataverse will continue to be software development, but in OpenDP I'll focus on community management. My guess is that to some extent, I'll work in both code and communities for both projects. I've been known to give a talk or two about Dataverse, but I'm fairly new to OpenDP. The “DP” is short for differential privacy.

Today is #ShutDownSTEM (or #ShutDownAcademia) and I spent a little time attending two webinars and hanging out in a Slack channel called #fighting-antiblackness. Someone linked to Police: Last Week Tonight with John Oliver. It was funny and interesting but I kept thinking that I should be listening to what black people have to say. Thankfully, at the very end, he played a speech by Kimberly Jones, author of I'm Not Dying with You Tonight.

Inkscape 1.0 was released today. The release video is absolutely fantastic and captures the energy of the Inkscape community. The announcement does a great job of describing new features with animations throughout. I eventually got a dev environment set up on my Mac and it was exciting to see the latest code running but I'm really not a C++ developer and haven't found much time to contribute. I did make a few contributions to the new Inkscape Beginner's Guide linked from the books section of the website.

Today I returned home after a long stay at McLean, a psychiatric hospital not far from Boston. I now recognize that I was getting increasingly manic in the days leading up to my trip to the emergency room and subsequent hospitalization. People at work and in my personal life kept asking me what had gotten into me. I kept wondering why they were moving, talking, and thinking so slowly. My “first break” (first mood episode) occurred on Thursday, March 5th.

This morning I was just starting my day when Oliver Bertuch, frequent Dataverse contributor and giver of gifts, pointed me to a test server making use of the OpenID Connect (OIDC) feature he added to Dataverse 4.19. As chronicled in today's IRC log, I captured numerous screenshots in a GitHub issue and expressed a lot of enthusiasm over how this next generation repository technology really works! During my one on one with my boss I explained how this opens up a lot of possibilities and could solve a variety of issues.

Today I couldn't remember some software I used 14 years ago to block IP addresses that are attacking your server in a denial of service (DoS) attack. I emailed a colleague and friend who was on the same team with me back then and he replied right away. Awesome. It turns out I did remember the software that he still uses, which is called fail2ban. The one I forgot is called DenyHosts, which he no longer uses.

Biking home from work today I was listening to the video of the second day of the NIH Generalist Repositories Workshop and heard mention of a spreadsheet being used to compare data repositories. Sure enough, the notes for that day mention a spreadsheet called Repository and Index Software (Work in Progress) as well as a repo on GitHub that was recently updated to include this spreadsheet. The repo says, “We created this list to:

Today I watched a short video about how Mozilla is moving from IRC to Matrix by Rubén Martín from Mozilla Reps. The slides are available as well. As far as I know, this is how the change unfolded: 2019-04-26 IRC will go away 2019-05-14 explanation of process 2019-09-06 four possibilities announced 2019-09-12 feedback gathering 2019-12-19 decision to go with Matrix announced 2019-12-19 welcome from Matrix 2020-02-20 we are live Now the Matrix page in the Mozilla wiki explains how to connect and the fact that federation is enabled.

With the exception of my first post and a short description of some editor and deployment mechanics, I've tried to avoid getting meta with this blog, but I thought I'd mention that I've settled on a “no blogging on vacation” policy. Eight days have passed since my last blog post. I had a nice streak going, blogging every day of 2020. I was staring to feel like Doogie Howser, typing away in the evening, reflecting on something that happened that day.

It's Valentine's Day which means it's Love Data Week. Lots of cute tweets in the air. In addition, the National Institutes of Health (NIH) hosted a workshop on “The Role of Generalist Repositories to Enhance Data Discoverability and Reuse,” which I learned about in a thread on the Dataverse mailing list. The video of day 1 is available and I really enjoyed the first keynote by Sayeed Choudhury, especially around 31:40 where he talked about the triad of academia, government, and the private sector:

Today I had an absolutely fabulous call with Ed Thomson and Rimas Silkaitis from GitHub. As I mentioned in my post about my talk at FOSDEM a couple weeks ago, Ed attended my talk and approached me afterwards saying he'd like to hear more the take Dataverse has on scientific reproducibility. Here are the notes from our meeting. It was so much fun talking to these guys. It makes me want to go back and listen to all the episodes of All Things Git, which I believe is where I first encountered Ed.

Today I popped into the Frictionless Data chat room to ask a quick question about a potential integration. This was a follow up on yesterday's Dataverse Community Call when Lilly Winfree joined (thank you!) to introduce herself and projects she's working on. I recently attended her talk called Frictionless Data for Reproducible Research and as I said on the call, I highly recommend watching the video. Her slides are great too.

Tonight I went to a fantastic talk by Amy X. Zhang called “Systems to Improve Online Discussion.” I met Amy last year at LibrePlanet and shortly afterward made a note to myself to follow up with her about communication platforms. Digging through old emails, it looks like I first reached out to her in June 2015. I head heard good things about Hackstack at the first Dataverse Community Meeting and read her paper called “Mailing Lists: Why Are They Still Here, What's Wrong With Them, and How Can We Fix Them?

Today I listened to a talk by Georgia Bullen called Beyond the Pile of Knobs: Usability and Design for Privacy, Security, Safety & Consent Privacy given in the Open Source Design devroom at FOSDEM. Her slides are available but the video is well worth listening to or watching. She advises, “Don't just focus on the majority cases, focus on the high-risk users and understand their threats.” She calls these the “stress cases,” the people who are in the most high-risk settings, and suggests that what works for them will likely work for most people.

My daughter was introduced to Desmos in her 8th grade math class and she's having a blast creating art with it. The Wikipedia article for Desmos describes it as an advanced graphing calculator and mentions toward the end that “another peculiar use of the calculator involves the creation of graphic arts.” Sure enough, the Staff Picks: Creative Art link on the Desmos home page reveals pictures of a shark, a chick, a seal, and more.

The first Inkscape Vectors meeting of 2020 was awesome and I returned for more! An issue is used to track the agenda. As usual, the meeting was logged. Here was the agenda, the stuff crossed out that wasn't talked about: Tutorials Funding @ryangorley @CRogers (10 min) Inkscope @tim (I will schedule a separate short meeting) 2020 Big Picture Goals @all (10 min) Revitalize Bug Hunt & bug migration (5 min) Inkscape Survey @victorwestmann (10 min) Social media stats update (NA) #131 website feature - highlight user project @brynn (10 min) building Documentation team, make announcement @brynn (5 min) Release (5 min) Tutorials Funding @ryangorley @CRogers (10 min) Abstract - Users on all major social media platforms have expressed a desire for instructional video offered by the project.

As planned, on Sunday I gave a talk at FOSDEM called “Advancing science with Dataverse: Publication, discovery, citation, and exploration of research data.” I was eager to introduce a bunch of open source hackers like myself to the related concepts of open science and open access. Afterward, Adina Wagner told me I did a good job of connecting these worlds. Phew. I'm also happy, of course, to get the word out about Dataverse.

The one talk I attended in person in the HPC devroom on Sunday at FOSDEM was Interactive HPC but now I'm working my way through the recordings. Towards reproducible Jupyter notebooks caught my eye and it was really interesting. The software is called Guix-Jupyter and I already mentioned it in a previous blog post where I pointed out a tweet that links to a post on the Jupyter forum. To motivate the discussion, we first see a tweet by Daniel Katz:

Today I told a couple members of the Sid team about a talk at FOSDEM I attended on Sunday called “Interactive applications on HPC systems Jupyterhub, Galaxy, RStudio, XPRA.” I pointed out that the slides show that they've looked at solutions like OnDemand but built their own thing. I'm tempted to put this new project on my list of things that look like Sid but I don't think it has a name and it isn't open source.

Today it was announced that Dataverse will apply to participate in Google Summer of Code 2020. Awesome. Update: I suggested a data driven approach of using Druthers to figure out what users want.

Today my friend Slava tweeted the following at me: “I'm thinking for a long time how to turn all SLOPI communication channels into #knowledgebase with some NLP pipelines and archive it in @dataverseorg. Probably another crazy idea for Hackathon in June during community meeting.” In my reply, I tried to explain that yes, I definitely want a knowledge base, but from a recent conversation with a friend, I'm wondering if it should be build on Blacklight.

I mentioned that my FOSDEM talk was accepted a while back. Awesome. Here are some devrooms I'm planning on checking out: Open Research Tools and Technologies devroom HPC, Big Data, and Data Science devroom Open Source Design devroom Free Java devroom video