Feeds

A few weeks ago (Time flies!) I attended the KDE Goals Sprint in Berlin. I didn't have concrete plans, but I intended to look into accessibility. Quite some time ago I had improved the accessibility of Kleopatra and at Akademy 2023 in Thessaloniki I gave a talk about it. Back then I had taken the easy route fixing everything directly in Kleopatra and working around several issues in Qt instead of fixing the issues in Qt itself so that all apps could profit. Time to do something about it.

(In-)Accessible icon-only buttons

A common problem for accessibility is icon-only buttons. If a button doesn't have text then screen readers can only tell their user that there's a button. That's not very helpful. Sometimes the developers have at least assigned a tool tip to the button. This can be read out by the screen readers (Qt provides the tool tip as accessible description of the button.), but it's often too verbose. To make a button without text accessible the developer has to set the accessible name property of the widget or, in case of a Qt Quick app, the name property of the Accessible QML Type. Unfortunately, that's often forgotten if the UI isn't designed with accessibility in mind.

At the sprint I discussed several ideas with other participants to help developers remember to set the accessible name:

• A helper class to instantiate in your app which inspects the app's windows and prints a report at the end with all inaccessible icon-only buttons it has found. A bit like Qt's QAbstractItemModelTester or the different compiler sanitizers. I implemented a prototype of such a class, but didn't pursue this further. The downside of this approach is that the developer needs to open each window of the app to find all inaccessible buttons. If they are already aware of the problem then it's probably easier to search the code.
• Instead of using a helper class to inspect the widget tree from the inside one could inspect the accessibility tree of the app from the outside. This could be built into our Appium-driven UI test framework so that developers don't have to do anything special. Except that they need to write UI tests that open each and every window of their app. I think it's still worth to look into this.
• Last but not least, we pondered writing a clazy test. Thinking about the many different ways a text can be set on a button (e.g. with KGuiItem::assign) we doubted that it would be feasible to write such a test.

In the end the easiest approach could be education. If the developers are aware of the problem then there's a good chance that they remember to set an accessible name the next time they add an icon-only button to their app.

Accessible date/time inputs

Volker, Carl, David, Harald and me discussed and explored some ideas to make the date and time inputs in Qt Quick apps like Itinerary accessible. In Kleopatra I resorted to allow the user to enter the date in a simple text input instead of trying to make the complex UI of KDE's date picker accessible. Read Volker's blog and David's blog to find out which solutions they found for Qt Quick apps.

A small automation interlude

One advantage of sitting with other people in the same room is that you may overhear them talking about a mistake (e.g. a faulty commit) and you know exactly how to prevent this kind of mistake in the future. In this case the problem was a missing quote character in some YAML file. And the preventive measure was adding a YAML linter CI job. While I was at it I removed some unnecessary code from the CI job and added the job to a second repository.

List views with underlying multi-column model

In several widgets that show or use a simple list of items Qt allows using a model with multiple columns, e.g. QListView, QComboBox, QCompleter. In general this works well except that Qt has a long-standing bug: When navigating through the list screen readers read the entries of the underlying model column by column instead of reading only the entries in the selected model column (QTBUG-33786). In Kleopatra I worked around this bug with a proxy model which pretended that the model only had one column.

During the sprint I finally sat down and prepared a fix for Qt. For better readability I split my changes in five separate commits which resulted in five separate patches for Qt: 556857, 556858, 556859, 556860, 556861. Being used to multi-commit MRs in GitLab I wondered if I had done something wrong when I submitted my changes, but apparently that's Gerrit's way of handling patch reviews. The first two commits are code clean-ups, the third commit changes one aspect of the accessibility test for list views, the fourth commit is the actual fix, and the fifth commit adds a few more unit tests I find useful. The first three commits have been merged, but the actual fix is still waiting for a review.

After that I looked into the problem that QListView emitted an accessibility focus event when the current item changed even if the list view didn't have focus. I found out that this had been fixed recently by a fellow Qt contributor who ran into the same problem independently of me. This meant that I could remove the workaround in Kleopatra for new enough Qt.

Thanks to MBition for hosting us and to Aleix for making sure we don't starve. And many thanks to those donating to KDE which makes these sprints possible.

Links and References For My PyCon US Keynote

It’s another year, and that means it’s time for another PyCon US. This event offers an opportunity to bring together a vibrant and diverse community of engineers, data scientists, researchers, students, and enthusiasts who use the Python programming language for work and for fun.

It also presents a unique opportunity to highlight entrepreneurs building the future of the Python ecosystem. 2024 marks the thirteenth year that PyCon US has invited early-stage companies to present their ventures on Startup Row. Tools and platforms like Pandas, Plotly, and more—now fixtures of the Python software ecosystem—were built by companies that were featured on Startup Row. And with each new year there’s an opportunity to catch a glimpse of what’s next.

PyCon US organizers and the Python Software Foundation are thrilled to welcome seven companies to Startup Row at PyCon US 2024.The Startup Row Lineup at PyCon US 2024DAGWorks

Alexander Hamilton and Aaron Burr might’ve been historic rivals, but the two open-source Python libraries named after them—Hamilton and Burr—work together quite nicely. The company behind Hamilton and Burr, DAGWorks, was founded with the goal of standardizing the way developers build data, LLM, and RAG pipelines, as well as managing and observing application state; both libraries also come with optional self-hosted UIs. Hamilton is the older of the two projects and is used in production by many companies, from banks to startups, to government departments, through to enterprise conglomerates. DAGWorks is based in the San Francisco Bay Area and was founded by Hamilton and Burr co-creators Stefan Krawczyk and Elijah ben Izzy, who previously worked together at Stitch Fix where they built the styling service’s MLOps stack. DAGWorks participated in Y Combinator’s Winter 2023 accelerator batch.

Dispatch

Building and maintaining distributed systems is a complex and often frustrating exercise. A given function might go rogue, either firing too frequently or silently failing to execute when needed. Dispatch is a startup aiming to make building reliable distributed systems a breeze. Its platform offers developers the ability to make any async function resumable by simply wrapping it with a @dispatch.function decorator, enabling durable workflows and robust error handling. Quoting a customer, ““I love the API design, just tuck in a decorator and you’re done, almost feels like cheating!” The company was founded by engineers with experience building high-scale distributed systems at companies like Meta, Segment, and Twilio, among others.

dltHub

Dealing with data can be a chore for developers. One has to actually get the data, clean it up, and that data has to get loaded somewhere for it to be useful. At least for that last step, there’s dlt (a.k.a. data load tool), an open source Python library that takes the hassle out of shepherding data from one place to another. The library features out of the box integrations for dozens of verified data sources and destinations, making it easy to fetch data from places like Facebook Ads, Pandas dataframes, or even Chess.com and then load data into destinations ranging from Databricks and Snowflake to Postgres and DuckDB. dltHub, the company behind the popular dlt library, is based in Berlin and New York City. “Today it is easier to pip install dlt and write a custom source in Python than to setup and configure a traditional ETL platform,” said CEO and co-founder Matthaus Krzykowski. The startup is backed by Dig Ventures and technical founders from companies like Huggingface, Miro, Matillion, and others.

Exaloop

What gets called “big” data is constantly shifting. Not too long ago, anything over a handful of gigabytes was pretty big. Now, working datasets can be several orders of magnitude larger, which presents a challenge for data scientists and developers who just want to write Python. Enter Exaloop. The MIT spinout company’s platform enables data scientists and engineers to write and run Python workflows that execute extremely fast (10x-100x faster than vanilla Python), even when dealing with truly massive data. The secret sauce: Codon, a high-performance Python compiler that uses LLVM. Codon has garnered a fair bit of attention from the developer community, racking up nearly 14,000 stars on Github. The company is planning to release a new entirely cloud-based platform, enabling users to leverage Exaloop’s tech and performance benefits conveniently within a browser window.

Pandas AI

Data analysis packages like Pandas are, to most users, pretty intuitive. Someone with a bit of Python knowledge can get the hang of these tools pretty quickly. But not everyone knows Python, and many others find the prospect of writing code intimidating. For those who prefer to ask questions rather than write complex queries, there’s PandasAI. PandasAI allows users to connect to a database or import a CSV, making data interaction conversational. “We want to make it easy for anyone, in any company, to be able to derive valuable insights from data. That shouldn’t be limited to a small team of engineers and analysts,” said PandasAI founder and CEO Gabriele Venturi. Users can write queries in natural language (e.g. “In my database of ecommerce transactions, which customer Id appears most often?”) and even command PandasAI to plot out the data they’re analyzing. The library is compatible with several LLM providers and model variants, reducing the risk of model lock-in. PandasAI is open source and has over 11,000 stars on Github. The company participated in Y Combinator’s Winter 2024 batch.

Pixee

When it comes to application security, many developers wish for the equivalent of magical fairy dust they can sprinkle on their codebase to fix bugs and vulnerabilities. Pixee is about as close as developers can get to that magical fairy dust. The D.C. metro area-based security startup is the maker of Pixeebot, a virtual security engineer that automatically triages and fixes potential security issues before they can be exploited in production. The bot is built on Codemodder, an open source framework created and maintained by Pixee, that can find and fix issues in Python and Java. Pixee intends to extend language support to JavaScript, and Node.js in the near future'

Trellis

There’s nothing better than clean, well-structured data. It’s easy to query, easy to work with, and easy to add value to any enterprise or personal use case. The problem is that data rarely comes so neatly packaged. Trellis helps developers ingest and extract insights from unstructured data sources like sales calls, regulatory filings, business contracts, large-scale web scrapes, and more. Trellis users can define their data schema in natural language, which is transformed into a structured SQL format that can integrate with hundreds of data sources and LLM-augmented workflows. The company is based in San Francisco and participated in Y Combinator’s Winter 2024 accelerator batch.

A Quick Look Back at 2023’s Startup Row Alumni

Startup founders make no small plans. Startups are inherently risky and not all are destined for greatness, but the 2023 batch of Startup Row companies seems to be on the path to success.

• 🦄 Imbue, a generative AI company developing agentic models that can reason and code, raised $200M+ in Series B funding at a $1B+ valuation. Investors in that deal include NVIDIA, Astera Institute, Notion co-founder Simon Last, and former Cruise CEO Kyle Vogt, among others. The company was formerly known as Generally Intelligent.
• Neptyne continued expanding its Python-included spreadsheet tools and launched a Google Sheets extension that brings real, live Python to Google’s spreadsheet software. Neptyne also launched API integrations, giving its users credits to call OpenAI, Bing web search, web page rendering through PhantomJsCloud, Google’s geocoder, and financial information within their spreadsheet environments.
• Nixtla released TimeGPT (ArXiv). Considering the company’s whole raison d'être is building time series data modeling and projection tools, this seems like a great use of transformer models’ next-token prediction capabilities. Nixtla’s open source tools continue to gain popularity.
• ❄️ Ponder Data, maker and maintainer of Modin, was acquired by Snowflake to expand the Python capabilities of its data cloud platform. Modin is used by hundreds of thousands of users and serves as a drop-in replacement for the ever popular Pandas library.
• Predibase. A month or so after PyCon 2023, the company announced that it had raised another $12.2 million from Felicis Ventures to bring total Series A funding up to $28M. In November the company released its LoRA Exchange (LoRAX) framework to the open source community under an Apache 2.0 license. According to Predibase’s announcement, “[the framework] makes it possible to serve hundreds of fine-tuned LLMs at the cost of one GPU with minimal degradation in throughput and latency.” In February the company launched LoRA Land, an LLM playground featuring 25 fine-tuned open source LLMs that “rival or outperform GPT-4” on specific tasks.
• Reflex is the new name of the company formerly known as Pynecone, and its open-source web application framework is garnering more attention than ever. Back in August the YC alumni company announced that it raised $5M in seed funding led by Lux Capital to continue building out its platform and develop and launch a revenue-generating hosting service for Reflex apps. The company’s open source framework has over 17,000 stars on Github.
• Wherobots, a geospatial data platform founded by the creators of Apache Sedona, announced $5.5 million in seed funding to build out its enterprise platform. Apache Sedona has been downloaded over 20 million times and is among the top one percent of most-downloaded packages on PyPI.

Needless to say, we’re pleased to be at least a small part of these companies’ ongoing success.

Thank You’s & AcknowledgementsStartup Row is made possible by the trust, logistical support, and tireless efforts of countless folks.
A heartfelt thanks goes out to the organizers and volunteers who work diligently to ensure PyCon US is a positive and engaging event, both in-person and virtually.
We also extend our gratitude to the many entrepreneurs who, despite their undoubtedly busy schedules, applied to Startup Row. Your time and attention are valuable, and we appreciate the thought you put into your applications.
A special thanks goes to the selection committee. Scoring such a competitive applicant pool is a significant commitment, and their time and expertise is greatly appreciated.
Lastly, our deepest thanks to the 7 companies that traveled to Pittsburgh to share their innovations with over 2,700 PyCon US attendees. Thank you all once again.

This is a brief overview of the project that I will be doing this summer as part of Google Summer of Code. As the title of this post says, I will create Python bindings for KDE Frameworks. There are more than 70 libraries, so the aim is to add support for three of them: KWidgetsAddons, KCoreAddons and KI18n. As predicting how much time we’ll need for each one is a little bit complex, I might end adding support to other libraries if the time allows.

Sunday, 19 May 2024

KDE today announces the release of KDE Frameworks 5.116.0.

KDE Frameworks are 83 addon libraries to Qt which provide a wide variety of commonly needed functionality in mature, peer reviewed and well tested libraries with friendly licensing terms. For an introduction see the KDE Frameworks release announcement.

This release is part of a series of planned monthly releases making improvements available to developers in a quick and predictable manner.

New in this version Breeze Icons

• Add audio/ogg and audio/x-vorbis+ogg icons
• Add audio/vnd.wave MIME type

Extra CMake Modules

• ECMAddQch: drop trying to set IMPORTED on targets with installed config
• Remove extraneous docs-build CI job that is no longer needed following the switch of api.kde.org to Gitlab CI

KActivitiesStats

• resultset: fix agent escape string

KCalendarCore

• Fix Calendar::updateNotebook event visibility updates

KContacts

• Restore country detection tests on FreeBSD
• Disable FreeBSD tests that recently started to fail in the CI

KDED

• Wait until kconf_update finished

KFileMetaData

• fix handling of attribute namespacing

KI18n

• KF5I18nMacros.cmake.in - don't look for python[2,3] on Windows
• KCountrySubdivision: unbreak support of iso-codes >= 4.16

KImageFormats

• TGA: added options support (bug 479612)
• More header checks (CCBUG: 479612) (bug 479612))

KIO

• Strip trailing slash in iconForStandardPath

KItemModels

• Trivial fix for crash in buddy() when sourceModel isn't set yet

KPackage Framework

• testpackage: Add a website so that the tests succeed

KRunner

• Add default arg to AbstractRunner QVariantList constructor

KService

• Fix warning: mimeType "x-scheme-handler/file" not found (bug 442721)

QQC2StyleBridge

• Localization support

Syntax Highlighting

• fix refs
• use (?:sub){0,2} to work with all pcre versions

Security information

The released code has been GPG-signed using the following key: pub rsa2048/58D0EE648A48B3BB 2016-09-05 David Faure faure@kde.org Primary key fingerprint: 53E6 B47B 45CE A3E0 D5B7 4577 58D0 EE64 8A48 B3BB

I don't like slow software. So I use profilers to make software faster. What I like even less, is slow profilers! And perf is sometimes slow for completely unavoidable reasons; to resolve source line information (needed primarily for figuring out inlining, at least in the default settings), you need to go ask libbfd. But libbfd comes from binutils, and binutils is GPLv3. And perf is part of the Linux kernel, which famously is GPLv2. So if you build perf against libbfd, the result is… nondistributable. Distros cannot ship them. Not Spiderman pointing at Spiderman, but Stallman pointing at Stallman. perf has to resort to calling out to addr2line over a pipe, which sometimes works well and sometimes… well, not. A couple of years ago, I suggested an improvement here that got me a small amount of attention, but it still isn't a really reliable way to do things.

But over the last 20 years, some other group has been busy making compilers and linkers and disassemblers and low-level binary stuff. And they were pretty careful to make their stuff GPLv2-compatible. So I give you… perf using libllvm for source line lookup (and disassembling).

Hoping for a constructive review process and that I can reach the 6.11 merge window :-)

New Addition to Education Case Studies - Italy

All Italian-Language Schools in South Tyrol Migrated to Free Software

I released a new Swift library, DotEnvy. It's a parser and loader for dotenv files.

Dotenv is a vaguely specified format that is supported by libraries found for most languages used in server-side development. The idea is that a twelve-factor app is supposed to read its configuration from environment variables, which can be a hassle to maintain during development. So you store them in a non-version controlled file called .env your application reads upon startup.

The format looks more or less like this:

KEY1=VALUE1 KEY2="VALUE2" KEY3="MULTILINE VALUE" KEY4='REFERENCE TO ${KEY3}'

The Swift libraries I could find seemed to lack features and had not seen updates in years. I don't think a library like this needs a huge number of features, but multiline strings and variable references were something I wanted. And writing parsers is fun.

DotEnvy has good test coverage and online documentation. There's also pretty good error reporting.

There's also a command line tool that allows you to syntax check a dotenv file or convert it to JSON. I was going add a launcher (i.e. run dotenv-tool launch sh and it'd export the environment from .env and run sh), but discovered that pseudo terminals are a pain and my Stevens has gone missing. Patches are welcome.

I accidentally used the same name as a Rust dotenv library, but I decided there's enough namespacing provided by the language support that the risk of confusion isn't too great.

fzf is one of my favorite shell tools. I have a ton of scripts where I use it for selection. Here's one for searching git history. git log -Gpattern allows you to search for commits that contain pattern in the patch text. Combine it with fzf and you get a pretty decent history search tool.

I have this saved as ~/bin/git-search-log, so I can invoke it as git search-log pattern or git search-log pattern branch:

#!/bin/bash set -euo pipefail # Ensure compatibility with fish etc by ensuring fzf uses bash for the preview command export SHELL=/bin/bash git log -G$@ --oneline | fzf \ --preview-window=bottom,80% \ --preview "echo {} | sed 's/ .*//g' | xargs git show --color" \ --bind 'enter:execute(commit=$(echo {} | sed "s/ .*//g") && git diff-tree --no-commit-id --name-only $commit -r | fzf --preview-window=bottom,80% --preview "git show --color $commit -- $(git rev-parse --show-toplevel)/\{}")'

When you run it you get an selection of matching commits, one per line, with a preview window showing the patch. If you hit enter on a commit, you get another fzf screen, this time allowing you to select files modified in that commit. Hit enter again and you're back in the first one.

In Python, "change" can mean two different things. Assignment changes which object a variable points to. Mutation, changes the object itself.

Table of contents

1. Remember: variables are pointers
2. Mutating a list
3. Mutation
4. Assignment
5. Assignments versus mutations
6. Changing variables and changing objects

Remember: variables are pointers

When talking about Python code, if I say we changed X, there are two different things that I might mean.

Let's say we have two variables that point to the same value:

>>> a = [2, 1, 3, 4] >>> b = a

Remember that variables in Python are pointers. That means that two variables can point to the same object. That's actually what we've done here.

Let's change the object that the variable b points to.

Mutating a list

If we append a number …

Read the full article: https://www.pythonmorsels.com/assignment-versus-mutation/

KDE’s KContacts framework provides API for locale-aware address formatting and address format metadata since quite some time, with an upcoming change this will all also be available for QML code directly.

Country-specific address formatting

Addresses are generally formatted differently depending on the country they are in. Such differences can be whether the state or region is relevant/included, how different parts are ordered or how different parts are joined in the local language/script.

If we have address information in a somewhat structured form, ie. broken up into individual parts (street, postal code, city, country, etc), displaying that correctly requires knowledge of those formatting rules. As this is not an uncommon problem, the KContacts framework provides C++ API for this. Using that from QML without custom glue code is now also becoming possible.

pragma ValueTypeBehavior: Addressable import org.kde.contacts import QtQuick.Controls Label { text: { const addr = { country: "DE", region: "BE", locality: "Berlin", postalCode: "10969", street: "Prinzenstraße 85 F" } as address; return addr.formatted(KContacts.AddressFormatStyle.MultiLineInternational, "KDE e.V."); } }

Different formatting styles are supported (single- or multi-line, international or domestic, for display or for postal mail labels).

Address format metadata

Additionally, the metadata necessary for formatting addresses can also be queried. This is useful for example for:

• Showing only the input fields in an address edit form actually relevant for a specific country.
• Input validation of postal codes, as shown in the code example below.
• Ordering input fields in the canonical order in a given country.

import org.kde.contacts import org.kde.kirigami import QtQuick.Controls import QtQuick.Layouts RowLayout { TextField { id: postalCodeEdit text: "SW1P 3EU" property string format: AddressFormatRepository.formatForCountry("GB", KContacts.AddressFormatScriptPreference.Local).postalCodeRegularExpression property bool isValid: text.match("^" + format + "$") } Icon { source: postalCodeEdit.isValid ? "dialog-ok" : "dialog-warning" } }

More elaborate examples can be found e.g. in Itinerary’s address/location editor.

What’s still missing

Review and approval of this MR.

I’ve just got a new Kogan 5120*2160 40″ curved monitor. It cost $599 including shipping etc which is much cheaper than the Dell monitor with similar specs selling for about $2500. For monitors with better than 4K resolution (by which I don’t mean 5K*1440) this is the cheapest option. The nearest competitors are the 27″ monitors that do 5120*2880 from Apple and some companies copying Apple’s specs. While 5120*2880 is a significantly better resolution than what I got it’s probably not going to help me at 27″ size.

I’ve had a Dell 32″ 4K monitor since the 1st of July 2022 [1]. It is a really good monitor and I had no complaints at all about it. It was clearer than the Samsung 27″ 4K monitor I used before it and I’m not sure how much of that is due to better display technology (the Samsung was from 2017) and how much was due to larger size. But larger size was definitely a significant factor.

I briefly owned a Phillips 43″ 4K monitor [2] and determined that a 43″ flat screen was definitely too big. At the time I thought that about 35″ would have been ideal but after a couple of years using a flat 32″ screen I think that 32″ is about the upper limit for a flat screen. This is the first curved monitor I’ve used but I’m already thinking that maybe 40″ is too big for a 21:9 aspect ratio even with a curved screen. Maybe if it was 4:4 or even 16:9 that would be ok. Otherwise the ideal for a curved screen for me would be something between about 36″ and 38″. Also 43″ is awkward to move around my desk. But this is still quite close to ideal.

The first system I tested this on was a work laptop, a Dell Latitude 7400 2in1. On the Dell dock that did 4K resolution and on a HDMI cable it did 1440p which was a disappointment as that laptop has talked to many 4K monitors at native resolution on the HDMI port with the same cable. This isn’t an impossible problem, as I work in the IT department I can just go through all the laptops in the store room until I find one that supports it. But the 2in1 is a very nice laptop, so I might even just keep using it in 4K resolution when WFH. The laptop in question is deemed an “executive” laptop so I have to wait another 2 years for the executives to get new laptops before I can get a newer 2in1.

On my regular desktop I had the problem of the display going off for a few seconds every minute or so and also occasionally giving a white flicker. That was using 5120*2160 with a DisplayPort switch as described in the blog post about the Dell 32″ monitor. When I ran it in 4K resolution with the DisplayPort switch from my desktop it was fine. I then used the DisplayPort cable that came with the monitor directly connecting the video card to the display and it was fine at 5120*2160 with 75Hz.

The monitor has the joystick thing that seems to have become some sort of standard for controlling modern monitors. It’s annoying that pressing it in powers it off. I think there should be a separate button for that. Also the UI in general made me wonder if one of the vendors of expensive monitors had paid whoever designed it to make the UI suck.

The monitor had a single dead pixel in the center of the screen about 1/4 the way down from the top when I started writing this post. Now it’s gone away which is a concern as I don’t know which pixels might have problems next or if the number of stuck pixels will increase. Also it would be good if there was a “dark mode” for the WordPress editor. I use dark mode wherever possible so I didn’t notice the dead pixel for several hours until I started writing this blog post.

I watched a movie on Netflix and it took the entire screen area, I don’t know if they are storing movies in 64:27 ratio or if the clipped the top and bottom, it was probably clipped but still looked OK. The monitor has different screen modes which make it look different, I can’t see much benefit to the different modes. The “standard” mode is what I usually use and it’s brighter and the “movie” mode seems OK for the one movie I’ve watched so far.

In other news BenQ has just announced a 3840*2560 28″ monitor specifically designed for programming [3]. This is the first time I’ve heard of a monitor with 3:2 ratio with modern resolution, we still aren’t at the 4:3 type ratio that we were used to when 640*480 was high resolution but it’s a definite step in the right direction. It’s also the only time I recall ever seeing a monitor advertised as being designed for programming. In the 80s there were home computers advertised as being computers for kids to program, but at that time it was either TV sets for monitors or monitors sold with computers. It was only after the IBM PC compatible market took off that having a choice of different monitors for one computer was a thing. In recent years monitors advertised as being for office use (meaning bright and expensive) have become common as are monitors designed for gamer use (meaning high refresh rate). But BenQ seems to be the first to advertise a monitor for the purpose of programming. They have a “desktop partition” feature (which could be software or hardware – the article doesn’t make it clear) to give some of the benefits of a tiled window manager to people who use OSs that don’t support such things. The BenQ monitor is a bit small for my taste, I don’t know if my vision is good enough to take advantage of 3840*2560 in a 28″ monitor nowadays. I think at least 32″ would be better. Google seems to be really into buying good monitors for their programmers, if every Google programmer got one of those BenQ monitors then that would be enough sales to make it worth-while for them.

I had hoped that we would have 6K monitors become affordable this year and 8K become less expensive than most cars. Maybe that won’t happen and we will instead have a wider range of products like the ultra wide monitor I just bought and the BenQ programmer’s monitor. If so I don’t think that will be a bad result.

Now the question is whether I can use this monitor for 2 years before finding something else that makes me want to upgrade. I can afford to spend the equivalent of a bit under $1/day on monitor upgrades.

• [1] https://etbe.coker.com.au/2023/06/06/dell-32-4k-displayport-switch/
• [2] https://etbe.coker.com.au/2022/06/29/philips-438p1-43-4k/
• [3] https://www.benq.com/en-us/monitor/programming/rd280u.html

Related posts:

1. Dell 32″ 4K Monitor and DisplayPort Switch After determining that the Philips 43″ monitor was too large...
2. Philips 438P1 43″ 4K Monitor I have just returned a Philips 438P1 43″ 4K Monitor...
3. cheap big TFT monitor I just received the latest Dell advert, they are offering...

A few weeks ago, some of us discovered that KDE apps just looked terrible when run in GNOME. A lengthy discussion on icon theming ensued, with various improvements made on both sides. The KDE effort was spearheaded by Christoph Cullmann, as already described in his post on the subject. In a nutshell, KDE apps opting into the new system that are run outside of Plasma will always have the Breeze style and icons available, unless overridden by the system or the user. Apps opting in so far include Kate, Konsole, and Dolphin. Feel free to help opt more apps in by using those commits as inspiration!

Dolphin itself also received a lot of special attention this week, in addition to other cool stuff:

New Features

Dolphin now gives you the option to enable previews for folders on remote locations. Be aware that this can cause slowdowns, and the UI tells you that, too (Sergey Katunin, Dolphin 24.08. Link):

Discover now handles the case where one of your Flatpak apps has been marked as “end of life” and replaced with another one; it gives you the opportunity to switch to the new one, or cancel and keep using the old one anyway (Harald Sitter, Plasma 6.1. Link):

Eagle-eyed readers have noticed this isn’t using the new dialog style. It hasn’t been ported yet. There are a lot of instances like this where Kirigami.OverlaySheet is inappropriately used as a confirmation dialog that need porting. UI Improvements

Dolphin’s ability to let you change things as root when kio-admin is installed has received a big upgrade: now it shows you a warning telling you what bad things you can do if you’re not careful, and also keeps a banner visible while you’re in root mode (Felix Ernst, Dolphin 24.08. Link):

Dolphin has received a number of UI improvements and better handling for viewing read-only folders (Jin Liu, Dolphin 24.08. Link)

Switched Spectacle over to using the common style for immutable toolview tabs in Kirigami apps (me: Nate Graham, Spectacle 24.08. Link):

KMenuEdit no longer annoyingly prompts you for confirmation when you delete a group (Kenny Hui, Plasma 6.1. Link)

The icons shown in our dialogs no longer themselves depict dialogs for mega dialog-ception; now they’re just normal colored icons (me: Nate Graham, Frameworks 6.3. Link):

Bug Fixes

Attempting to open multiple “New Folder” dialogs on a slow network location no longer causes Dolphin to crash (Akseli Lahtinen, Dolphin 24.08. Link)

Very small SVG images are now displayed properly in thumbnail previews (Méven Car, kio-extras 24.08. Link)

Fixed a case where our authentication system could crash and leave apps unable to request authentication (me: Nate Graham, Plasma 6.0.5. Link)

Turning on HDR mode no longer makes the screen colors wrong when using Night Color (Xaver Hugl, Plasma 6.0.5. Link)

Screens using fractional scale factors no longer get a weird row of pixels on the bottom edge that are held to the color of previously opened windows (Xaver Hugl, Plasma 6.0.5. Link)

Fixed several Plasma crashes that were introduced by porting some custom drag-and-drop code to the upstream Qt thing, but turns out to not be suitable for our purposes. Reverting to our custom code fixes the crashes (Kai Uwe Broulik, Plasma 6.1, Link 1 and link 2)

When Chromium-based browsers are running in native Wayland mode, dragging and dropping files into websites no longer makes them freeze and crash. This was a complicated bug largely caused by Chromium doing something unusual, but KWin now handles it properly (David Edmundson, Plasma 6.1. Link)

Visiting System Settings’ File Search page no longer sometimes causes a long hang when the file indexer is under heavy load (Janet Blackquill, Frameworks 6.3. Link)

If for some reason you want to use the Kickoff application launcher to search for a single character, close Kickoff, and then do the same thing again, the second search will now show results as expected (Alexander Lohnau, Frameworks 6.3. Link)

KSvg items and Kirigami.Icon used in Plasma now re-color re-colorable SVG images in the expected way when they’re displayed in Plasma irrespective of color scheme. This makes the CatWalk cat look correct while using a mixed light/dark global theme like Breeze Twilight (Marco Martin, Frameworks 6.3. Link):

Other bug information of note:

• 3 Very high priority Plasma bugs (same as last week). Current list of bugs
• 39 15-minute Plasma bugs (up from 38 last week). Current list of bugs
• 106 KDE bugs of all kinds fixed over the last week Full list of bugs

…And Everything Else

This blog only covers the tip of the iceberg! If you’re hungry for more, check out https://planet.kde.org, where you can find more news from other KDE contributors.

How You Can Help

The KDE organization has become important in the world, and your time and labor have helped to bring it there! But as we grow, it’s going to be equally important that this stream of labor be made sustainable, which primarily means paying for it. Right now the vast majority of KDE runs on labor not paid for by KDE e.V. (the nonprofit foundation behind KDE, of which I am a board member), and that’s a problem. We’ve taken steps to change this with paid technical contractors—but those steps are small due to growing but still limited financial resources. If you’d like to help change that, consider donating today!

Otherwise, visit https://community.kde.org/Get_Involved to discover other ways to be part of a project that really matters. Each contributor makes a huge difference in KDE; you are not a number or a cog in a machine! You don’t have to already be a programmer, either. I wasn’t when I got started. Try it, you’ll like it! We don’t bite!

 I've been on Chromebooks for a while.  However, since I had to recently try a Mac, I figured it was time to give Firefox a try again.  After two weeks of trying, I've given up.  At least for myself, I figured I'd write down the reasons I've given up.

Reasons:

• There is no way from the tab context menu to move a tab between windows.  I typically try to keep no more than 3 windows open at a time.  Ideally one, but maybe a second.  Without the ability to through the context menu to move a tab, I need a very large screen (not a laptop screen) to move tabs between windows.
• I couldn't find a way to take a URL and turn it into a custom search.  This really is a critical feature as it allows me to use short names to access specific searches.  E.g. search code (cs), show a calendar (c), etc.

De 27 a 30 de abril de 2024 foi realizada a MiniDebConf Belo Horizonte 2024 no Campus Pampulha da UFMG - Universidade Federal de Minas Gerais, em Belo Horizonte - MG.

Esta foi a quinta vez que uma MiniDebConf (como um evento presencial exclusivo sobre Debian) aconteceu no Brasil. As edições anteriores foram em Curitiba (2016, 2017, e 2018), e em Brasília 2023. Tivemos outras edições de MiniDebConfs realizadas dentro de eventos de Software Livre como o FISL e a Latinoware, e outros eventos online. Veja o nosso histórico de eventos.

Paralelamente à MiniDebConf, no dia 27 (sábado) aconteceu o FLISOL - Festival Latino-americano de Instalação de Software Livre, maior evento da América Latina de divulgação de Software Livre realizado desde o ano de 2005 simultaneamente em várias cidades.

A MiniDebConf Belo Horizonte 2024 foi um sucesso (assim como as edições anteriores) graças à participação de todos(as), independentemente do nível de conhecimento sobre o Debian. Valorizamos a presença tanto dos(as) usuários(as) iniciantes que estão se familiarizando com o sistema quanto dos(as) desenvolvedores(as) oficiais do projeto. O espírito de acolhimento e colaboração esteve presente em todos os momentos.

Números da edição 2024

Durante os quatro dias de evento aconteceram diversas atividades para todos os níveis de usuários(as) e colaboradores(as) do projeto Debian. A programação oficial foi composta de:

• 06 salas em paralelo no sábado;
• 02 auditórios em paralelo na segunda e terça-feira;
• 30 palestras/BoFs de todos os níveis;
• 05 oficinas para atividades do tipo mão na massa;
• 09 lightning talks sobre temas diversos;
• 01 performance Live Eletronics com Software Livre;
• Install fest para instalar Debian nos notebook dos(as) participantes;
• BSP (Bug Squashing Party - Festa de Caça à Bugs);
• Uploads de pacotes novos ou atualizados.

Os números finais da MiniDebConf Belo Horizonte 2024 monstram que tivemos um recorde!

• Total de pessoas inscritas: 399
• Total de pessoas presentes: 224

Dos 224 participantes, 15 eram contribuidores(as) oficiais brasileiros sendo 10 DDs (Debian Developers) e 05 (Debian Maintainers), além de diversos(as) contribuidores(as) não oficiais.

A organização foi realizada por 14 pessoas que começaram a trabalhar ainda no final de 2023, entre elas o Loïc Cerf do Departamento de Computação que viabilizou o evento na UFMG, e 37 voluntários(as) que ajudaram durante o evento.

Como a MiniDebConf foi realizado nas instalações da UFMG, tivemos a ajuda de mais de 10 funcionários da Universidade.

Veja a lista com os nomes das pessoas que ajudaram de alguma forma na realização da MiniDebConf Belo Horizonte 2024.

A diferença entre o número de pessoas inscritas e o número de pessoas presentes provavelmente se explica pelo fato de não haver cobrança de inscrição, então se a pessoa desistir de ir ao evento ela não terá prejuízo financeiro.

A edição 2024 da MiniDebconf Curitiba foi realmente grandiosa e mostra o resultado dos constantes esforços realizados ao longo dos últimos anos para atrair mais colaboradores(as) para a comunidade Debian no Brasil. A cada edição os números só aumentam, com mais participantes, mais atividades, mais salas, e mais patrocinadores/apoiadores.

Atividades

A programação da MiniDebConf foi intensa e diversificada. Nos dias 27, 29 e 30 (sábado, segunda e terça-feira) tivemos palestras, debates, oficinas e muitas atividades práticas. Já no dia 28 (domingo), ocorreu o Day Trip, um dia dedicado a passeios pela cidade. Pela manhã saímos do hotel e fomos, em um ônibus fretado, para o Mercado Central de Belo Horizonte. O pessoal aproveitou para comprar várias coisas como queijos, doces, e lembrancinhas, além de experimentar algumas comidas locais.

Depois de 2 horas de passeio pelo Mercado, voltamos para o ônibus e pegamos a estrada para almoçarmos em um restaurante de comida típica mineira.

Com todos bem alimentados, voltamos para Belo Horizonte para visitarmos o principal ponto turístico da cidade: a Lagoa da Pampulha e a Capela São Francisco de Assis, mais conhecida como Igrejinha da Pampulha.

Voltamos para o hotel e o dia terminou no hacker space que montamos na sala de eventos para o pessoal conversar, empacotar, e comer umas pizzas.

Bolsas para participantes

em breve

Financiamento coletivo

Pela segunda vez, fizemos uma vaquinha e conseguimos recursos para pagar algumas coisas (café da manhã, local, imprensa, crachá, cordões). A meta do crowdfunding era de R$ 3.000,00 e conseguimos R$ 3.940,00, recebemos de 62 doadores do Brasil e de outros países.

Cada participante recebeu: crachá, cordão personalizado, flyer com dicas "como contribuir com o Projeto Debian" e flyers de nossos patrocinadores.

Fotos e vídeos

Você pode assistir as gravações das palestras nos links abaixo:

• Youtube
• Peertube
• video.debian.net

E ver as fotos feitas por vários(as) colaboradores(as) nos links abaixo:

• Google photos
• Nextcloud

Agradecimentos

Gostaríamos de agradecer a todos(as) os(as) participantes, organizadores(as), voluntários(as), patrocinadores(as) e apoiadores(as) que contribuíram para o sucesso da MiniDebConf Belo Horizonte 2024.

Patrocinadores

Ouro:

• Collabora

Prata:

• Jedai.ai
• Toradex Brasil
• Globo.com
• Policorp Tecnologia

Bronze:

• BRDSoft - Tecnologias para TI e Telecomunicações
• EITA - Cooperativa de Trabalho Educação, Informação e Tecnologia para Autogestão

Apoiadores

• ICTL - Instituto para Conservação de Tecnologias Livres
• Nazinha Alimentos
• DACOMPSI

Organização

• Projeto Debian
• Comunidade Debian Brasil
• Comunidade Debian MG
• DCC/UFMG - Departamento de Ciência da Computação da Universidade Federal de Minas Gerais

Model UX Research & Design Docs for Command-Line Open Source

Celebrate Beautiful Soup's 20th Anniversary

A Celebration of My Friend, Dr. Mel Chua

User Support, Equanimity, And Potential Cross-Project Tools and Practices in Open Source

Maintainer Burnout: PyCon US 2023 Followup