May 2023 Summaries

Felt introduces an innovative commenting feature that allows users to collaborate directly on maps, collecting feedback in real-time and in context without the need for a Felt account. This feature aims to streamline the traditional feedback process, which often relies on indirect methods such as emails and meetings, by facilitating direct communication on the map itself. Users can engage in threaded conversations, tag teammates, and receive notifications about new comments, enhancing the overall collaboration experience. The new storytelling feature, "groups," allows elements to be organized within the map legend for improved viewing. Felt's entirely web-based platform is designed for seamless communication, making it ideal for teams, clients, and public interactions, and distinguishes itself by providing a modern, efficient way to work on maps collaboratively.

Felt introduces a modern, intuitive commenting feature for map collaboration that allows users to provide real-time feedback directly on maps, streamlining communication and enhancing the collaborative process. This feature eliminates the need for traditional, complex feedback methods like emails or meetings by enabling comments to be made in context, without requiring a Felt account. Users can invite others to view and comment on maps, with options to tag teammates, resolve conversations, and react with emojis, all of which are accessible from any device. Additionally, Felt includes a new storytelling feature, groups, which allows map elements to be organized in the legend for a better viewing experience. This web-based tool aims to make map-making a seamless, engaging, and efficient process for teams, clients, and public interactions, setting a new standard for collaborative map work.

In the transition from SVG to Canvas for rendering in Felt, the team not only aimed to enhance performance but also sought to improve their interaction-handling code. The new system was designed to address issues with the prior React and SVG approach, which included performance limitations, interactivity issues, and maintainability challenges. The newly devised interaction system, based on per-state InteractionHandlers, decouples interactions from rendering and access control, allowing interactions to be more performant and maintainable. This system enables precise and complex interactions by centralizing interaction management, which prevents overlapping or conflicting interactions and simplifies the code structure. The new approach allows for seamless interactions such as dragging elements through others without interference, a task previously unattainable with the element-centric model. Although the transition required overcoming challenges such as spatial indexing and performance optimization, the resulting system provides a more manageable and efficient framework that prioritizes user-requested features, enhancing the overall map-making experience in Felt.

The transition from SVG to Canvas in rendering elements enabled improved performance and interaction handling in Felt, a mapping application. Initially, SVG and React were used, but complexities arose due to performance issues and interactivity limitations, particularly when managing element interactions and maintainability. A new interaction system was developed, decoupling interactions from rendering and each other, allowing for more efficient handling of complex interactions without being tied to React or a specific rendering system. This system introduced a behavior-centric approach, using per-state interaction handlers, which simplified the architecture and improved performance by reducing resource allocation. The new method allowed for seamless interactions such as dragging elements through unselected items while still enabling selection, which was previously challenging with a DOM-centric approach. Despite some trade-offs, like adjusting to an app-wide interaction context and optimizing browser-native performance, the new system offers a cleaner, more maintainable codebase. This change aims to enhance the user experience by allowing developers to focus more on feature development and less on handling intricate interaction bugs.

Westchester Land Trust (WLT), established in 1988 in Bedford, New York, is dedicated to land protection through mechanisms like land acquisition and conservation easements to combat climate change. The organization, among the first to meet high professional standards as an accredited land trust, has Gentian Falstrom as its Land Project Manager. Despite lacking prior GIS experience, Falstrom now creates maps daily to prioritize land protection, serving as the initial contact for those interested in collaborating with WLT. Her role involves assessing environmental benefits using county, state, and conservancy datasets, scoring parcels based on their conservation values, and presenting these to the board of directors. A significant endeavor she leads is the creation of a public web map for WLT's new website, showcasing conserved lands while balancing public access with privacy for private landowners, using Felt as a user-friendly tool to achieve this without a developer.

Westchester Land Trust (WLT), established in 1988 in Bedford, New York, is committed to combating climate change through land protection strategies such as land acquisition and conservation easements. As one of the first accredited land trusts in the U.S., WLT upholds high professional standards. Gentian Falstrom, the Land Project Manager, plays a crucial role in this mission by using GIS mapping to prioritize land protection projects, despite having no prior experience with GIS before joining WLT. She evaluates potential projects by assigning scores based on environmental benefits, using data from county and state sources, as well as The Nature Conservancy. Recently, she has been working on a public web map for WLT's new website to visualize the impact of their conservation efforts, which presents unique challenges due to the variety of land types managed by WLT and the need to protect the privacy of certain landowners. To address these challenges, WLT has utilized Felt, a user-friendly mapping tool, to create the web map without needing a developer.

Felt transitioned from using SVGs with React to Canvas for rendering interactive maps to address performance challenges associated with rendering thousands of elements simultaneously. This shift, detailed in a two-part series, focused initially on rendering and performance improvements. Felt encountered limitations with SVGs, such as slow rendering during map interactions, noticeable delays, and visual artifacts, especially on large maps. The new Canvas approach redraws everything from scratch each frame, providing a more consistent performance profile and better user experience. Implementing Canvas involved feature flags, end-to-end tests, and performance benchmarks to ensure a seamless transition, which ultimately resulted in enhanced responsiveness, particularly on lower-end devices. The company employed various optimizations, such as caching text as textures and clipping off-screen elements, to boost performance. This switch was part of a comprehensive overhaul aimed at delivering the best user experience without requiring major visible changes for users, leading to improved map interaction and speed.

Felt transitioned from using SVGs with React to Canvas for rendering interactive maps to address performance challenges associated with rendering thousands of elements simultaneously. The move was driven by the need to improve speed and smoothness, particularly when handling complex maps with tasks like zooming and panning. SVGs required React to frequently create and update elements, which led to delays and visual artifacts. Canvas, on the other hand, redraws everything from scratch each frame, providing a stable frame rate and resolving many performance issues. Felt employed several optimizations, such as caching projected shapes and rendering lower-detail versions of elements at low zoom levels, to enhance Canvas performance. This transition was supported by comprehensive end-to-end testing, feature flags, and performance benchmarks, ensuring a seamless switch that resulted in faster and more responsive maps, especially on lower-end devices and mobile platforms.

Phoenix applications benefit from the built-in capabilities of the BEAM for long-lived processes, but a Graceful Shutdown is essential for cleaning up when those processes end, especially when additional processes are involved. In the context of a Phoenix application, a Graceful Shutdown involves stopping incoming work, completing ongoing work, and safely bringing down the system without disrupting the user experience, particularly during frequent code deployments. This process is supported by Phoenix's default configuration, including components like Plug.Cowboy.Drainer, which manages HTTP connection closures. However, challenges can arise with clustered Phoenix applications, such as maintaining accurate presence information. Implementing strategies such as Phoenix.Tracker.graceful_permdown helps ensure the system accurately reflects node status, preventing outdated presence data. Properly managing the application lifecycle, including the order of process termination and utilizing health checks, ensures a smooth transition during deployments, allowing for zero-downtime and a seamless user experience.

Graceful Shutdown is a crucial process in managing Phoenix applications, particularly for ensuring a seamless transition when old servers are replaced with new ones. This practice involves stopping new work acceptance, completing in-progress tasks, and safely bringing the application down without causing disruptions to users, essential for companies like Felt that deploy new code multiple times a day. Phoenix handles most shutdown processes through its default configuration, notably using Plug.Cowboy.Drainer to manage HTTP connections and automatic client reconnections to new servers. Effective system design for Graceful Shutdown includes structuring your application to avoid cyclic dependencies and ensuring readiness checks via endpoints. Properly handling shutdowns can prevent errors and maintain accurate presence information, especially in clustered environments, by using techniques like calling Phoenix.Tracker.graceful_permdown. Testing and implementing these strategies ensure that users experience minimal disruption during deployments, maintaining application reliability and performance.

Various individuals and groups have utilized the Felt mapping platform to create diverse and innovative maps for different purposes, showcasing its versatility and ease of use. Notable projects include Chris Holmes' map of Walvis Bay, which integrates vector layers over raster imagery, and Angelica Rodriguez-Arriaga's educational map on California's water history, enhanced with LiDAR data. Other creative uses include Chris Soriano's historical comparison map of Paterson, New Jersey, Mesut Ersöz's viral map of 90s music video locations in Istanbul, and a planning map for Santa Cruz's West Cliff Recreation Area developed by Stamen Design and Al Ramadan. Additionally, Katherine Rudzki highlighted pedestrian improvements in Raleigh, while Lanrick Bennett Jr. focused on mobility upgrades in Toronto. Ian Wolfe mapped multi-unit housing projects in Baltimore, Arun Ganesh created an interactive electoral ward map for Mumbai, and Márton Leitold designed a biking network around Hungarian hot springs. Felt users are encouraged to share their creations in a dedicated Slack group to gain recognition.

Felt maps are being utilized in a variety of innovative ways by diverse professionals, ranging from urban planners to educators, to illustrate complex data and support community projects. Chris Holmes used Felt to layer vector data over satellite imagery of Walvis Bay, while Angelica Rodriguez-Arriaga incorporated LiDAR data to teach California water history. GIS analyst Chris Soriano compared historical and current maps of Paterson, New Jersey, revealing urban changes, and Mesut Ersöz crafted a viral map of Istanbul's iconic 90s music video locations. Environmental concerns in Santa Cruz were highlighted through a map by Stamen Design and Al Ramadan, while Katherine Rudzki and Lanrick Bennett Jr. focused on pedestrian improvements and mobility upgrades in Raleigh and Toronto, respectively. Ian Wolfe showcased multi-unit housing developments in Baltimore, and Arun Ganesh provided an interactive electoral ward map for Mumbai. Márton Leitold's map connected Hungarian thermal waters with cycling routes, demonstrating the versatility of Felt in visualizing geographic and thematic data.