Skip to content
#small language model Open access

aipoch/open-science: v0.30.1

Sep 2026 · Zenodo (CERN European Organization for Nuclear Research)

Abstract

✨ Highlights Review skill updates as a real diff. Update previews render through a shared source diff viewer — file headers with change counts, line-number gutter, dashed red deletion and solid green addition rails, optional syntax highlighting, and readable fallbacks for malformed or oversized patches. (#2632) Side chats start as drafts with the right model. New side chats open an empty draft immediately, inherit the main conversation's current model and reasoning effort, and apply your own selection on the next send; untouched empty drafts are discarded automatically. (#2598) A standalone reset utility for Windows. When damaged data survives reinstalling, a guided command-line tool lists resolved data, configuration, profile, and authenticated runtime caches, refuses unsafe targets, and deletes only after you type an exact confirmation — with running runtimes blocking cleanup. (#2626) Small conveniences across the workspace. Tag and resource pickers gain keyboard search and creation, the project license is shown during installation, .science packages can carry the literature PDFs you select, and run previews and buttons share one motion language. (#2600, #2591, #2637, #2639) 🚀 New Features Shared highlighted diff viewer — reusable across the app and first wired into the skill update dialog, with translated omitted-context separators and screen-reader labels that keep the change meaning. (#2632) Side-chat drafts and conversation models — drafts with text or annotations survive navigation, provider-default intent survives restarts and resume failures, and the send menu no longer looks unavailable with an empty composer. (#2598) Standalone Windows data reset utility — PowerShell-based, linked from documentation; it validates every target before deletion, unlinks descendant links without following them, processes configuration last, and keeps errors visible. (#2626) Tag pickers with keyboard search and creation across resource pickers (#2600), license presentation during installation (#2591), selectable literature PDFs in .science packages, an animated shared run preview card (#2637), and unified button motion and action feedback (#2639). 🔧 Improvements Incompatible providers are probed on their own route instead of disturbing the active one, and keyless local gateways are now allowed. (#2569) Web reading approval is remembered for the rest of the conversation instead of being asked again. (#2635) PDF evidence notes clarify their limitations and diagnose bookmark sources. (#2594) The skill marketplace resolves update conflicts with reviewed in-place updates, and specialist release asset redirects are followed safely. (#2615) 🐛 Bug Fixes Sessions and plans — session-plan approval blocking survives MCP timeouts (#2631); verified historical artifact bindings are restored (#2633); task execution cancellation synchronizes with session admission (#2618); runtimes are preserved when session startup is still pending (#2620); side chats are excluded from package export scope (#2619) and OpenCode side-chat instructions no longer leak into main conversations (#2624); literature views return to the originating project conversation (#2599). Notebook and compute — queued runs and runtime state stay consistent (#2589); cleanup retries retain native termination proof (#2623); inaccessible inherited PATH directories are tolerated (#2613); already-hidden Linux paths are not masked again. PDF and previews — blocked parsing is recovered and extraction failures are reported (#2597); native figure and table layouts are recovered again after regression (#2617); background overlays stay below active modals. (#2593) Skills and connectors — ESM-2 batch residue pooling is corrected (#2601); gnomAD region arguments are validated (#2616) and reference builds are matched to datasets (#2596). Platform — macOS read-only installations are guided to fix permissions before updating (#2603); missing file-backed Codex credential imports are explained in Settings (#2609); tag assignment feedback is immediate again (#2610) and the pointer cursor stays while saving.

View source

Similar papers

#small language model Dataset Open access Oct 2026

Socratic guiding questions in synthetic arithmetic data: matched LoRA runs (revision v2)

Supporting data, adapters, predictions and code for the article *Low-Cost LoRA Fine-Tuning of Small Language Models for Multi-Step Arithmetic Reasoning* by Jake O'Grady, Asena Isik Gürhan, Chee Fong Ting and Effirul Ramlan (University of Galway). We generated 20,000 GSM8K-derived arithmetic problems with step-by-step s...

O'Grady, Jake, Gürhan, Asena Isik, Chee, Fong Ting et al. · 465 citations
#computer vision Open access Jun 2016

Software Development in Startup Companies: The Greenfield Startup Model

The results are packaged in the Greenfield Startup Model (GSM), which explains the priority of startups to release the product as quickly as possible, and the need to shorten time-to-market, by speeding up the development through low-precision engineering activities.

Carmine Giardino, Nicolò Paternoster, M. Unterkalmsteiner et al. · 178 citations · ⚡14
#computer vision Open access Oct 2016

Software Startups - A Research Agenda

Software startup companies develop innovative, software-intensive products within limited timeframes and with few resources, searching for sustainable and scalable business models.

M. Unterkalmsteiner, P. Abrahamsson, Xiaofeng Wang et al. · 157 citations · ⚡17
#machine learning Review Open access Oct 2016

“Failures” to be celebrated: an analysis of major pivots of software startups

This study conducts a case survey study based on the secondary data of the major pivots happened in 49 software startups, and demonstrates that customer need pivot is the most common among all pivot types.

Sohaib Shahid Bajwa, Xiaofeng Wang, Anh Nguyen-Duc et al. · 127 citations · ⚡15
#computer vision Review Open access May 2015

A survey study on major technical barriers affecting the decision to adopt cloud services

The comparison of adopter and non-adopter sample reveals three potential adoption inhibitor, security, data privacy, and portability, which underlines the importance of the technical and security perspectives for research investigating the adoption of technology.

Nattakarn Phaphoom, Xiaofeng Wang, S. Samuel et al. · 111 citations · ⚡8
#computer vision Open access Feb 2018

Lean Internal Startups for Software Product Innovation in Large Companies: Enablers and Inhibitors

This study investigates how Lean internal startup facilitates software product innovation in large companies and identifies its enablers and inhibitors, and shows the potential of the method-in-action framework to investigate the Lean startup approach in non-startup context.

Henry Edison, Nina M. Smørsgård, Xiaofeng Wang et al. · 78 citations · ⚡6

Related blog posts

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.