Invitation to participate into a "GMT Advisory Group" meeting

Announcements
1

Hello,

In the process of migrating the GMT infrastructure to EarthScope we are having a set of regular monthly meetings. Next one will be tomorrow at 19:00 UTC (= GMT :smile:) time.

We would like to invite interested people to participate in this meeting where things related to future of GMT are discussed. Sorry to announce this so close to the event but we are still adapting to a politics of opening to a wider range of participants.

The meeting will take place through a Zoom session, and as a first experiment we are being cautious not to post the Zoom link directly here on open because abuses from 3rd parties are a real risk. So, if you are interested in participating (even if only for assisting), please manifest your interest here by sending a contact email where to we can send the Zoom link, or to me directly (via forum message for example).

Hope to see new faces tomorrow

Joaquim

1 Like
2

Will be on the road at that time, but I like the initiative :slight_smile:

3

If you’ll not be driving, mobiles are just fine for this purpose.

4

Will the recording be made available?

5

We have not been recording the meetings (I think).

6

What was the meeting about? Maybe a list of 3-5-10 most important points, if possible? I’m curious.

People interested in and capable of developing a complex project like gmt seems to be quite a narrow range to me.

7

This was the agenda and other points that raided up during discussion.

GMT Advisory Meeting Agenda

March 12, 2026

  1. Update on transfer of GMT Forum and Remote Datasets to EarthScope/AWS infrastructure
    1. Need to establish landing area in ‘the Cloud’ for Discourse, consulting with IT on SaaS option.
      Remote Datasets likely to go to S3, with a review of how access is to be afforded. We will need access to the machine with those datasets. Need an account to manage those files by a couple of people. Develop plan for maintaining the datasets. SSH access paradigm on a server with GMT paradigm likely a problem, consider other options for updating and managing the datasets.
    2. Give Solar access to the current dataserver to assist in the process of migrating the data.
    3. Working on using a discourse hosted server, combining the GMT service with other Earthscope user communities.
    4. Explore how GMT is doing upload and download interface and how S3 might be brought in
    5. Would eventually need to just change the environment variable with a new html address. And updating default server when appropriate, would be a ‘breaking’ change.
    6. Would want new uploads to get replicated on mirror servers, will need to accomodate
    7. Timing, can data be transferred by the end of May?
    8. There are test failures due to network failures in the remote data transfer, it fails sometime, but then works the next time. Maybe moving to S3 would be more reliable.
  2. GMT Summit plans
  3. Put out the question of releasing GMT, 6.7
    1. Looking for the last half of April, 2026
    2. What needs to be done?
    3. Joaquim needs to fix something in the GMT installer for Windows
    4. The GMT code builds the documentation based on some cmake variables.
    5. Further discussion in the next meeting.
  4. Ideas for attracting and supporting contributors
    1. Inclusion of Asian community
    2. Somehow supporting AI access to assist contributors?
    3. Documentation for GMT: In general, it’s pretty good, but it can be hard to use. AI can help, but the core docs could use some cleanup.
    4. Expanding Advisory group?
      1. Seeking broader ideas and concerns for inspiration
      2. Avoid being a ‘closed shop’
      3. How big is too big for effective management
  5. Other ideas?
    1. Christian is interested in creating group collaborating on the PyGMT to perpetuate the system into the future,
    2. The use of the postscript library could become a problem should think about an alternative in case the postscript library becomes
    3. bgr.bund.de is a German organization is a heavy user of GMT in their presentations. Maybe a resource for expanding the user community.
    4. Explore finding an Asian friendly time for the GMT advisory meeting to occasionally meet.
    5. Biggest area for gmt development can be done through the wrappers, like in Julia or in PyGMT.
1 Like
8

Hi Joaquim. Can you please include me in future notifications regarding these meetings. Hopefully I can attend or my colleague Nathan who is our GMT magician. Thanks. Jonathan

Jonathan.Weiss@noaa.gov

9

Sure Jonathan,

Next meeting should be at 12 April but it is @SolarEarth who’s been leading and providing us the zoom link.

10

Do you guys mean AI is good enough answering questions about what gmt options are needed for solving some specific task if the user is having hard time reading the docs?

11

The next meeting is scheduled for 16 April. I’ve added your email to the invitation list.

12

AI can be useful for users in getting the options set up. But we were also meaning it can be very useful for developers or contributors hunting down bugs or making improvements to the software.

13

Zoom generated an overall summary and ‘chapter’ summaries from the recorded video.
I collected those summaries and pasted them here.

AdvisoryMeetingSummary

The GMT development team met to discuss the transition of data storage to AWS S3, with Garrett and Joaquim explaining that moving to S3 would improve download speeds and stability, particularly for users in Europe. The team explored the technical aspects of adapting GMT’s current data retrieval methods to work with S3 URLs, with Solar taking responsibility for investigating the implementation details. Christian, a long-time GMT user from Germany, shared his experience using GMT at BGR and expressed interest in potentially contributing to GMT’s development, particularly regarding documentation. The team also discussed the upcoming GMT 6.7 release planned for April, with Joaquim explaining his plans to update the Windows installer and address GPS-related issues. The conversation ended with a discussion about expanding the GMT community by including more international participants, particularly from Asia, in future meetings.

From Zoom Smart Chapters:

AWS S3 Data Transfer Planning

Rob discussed the transfer of gridded data to AWS S3, emphasizing the need to locate and transfer the data in its raw form. Garrett approved Rob’s plan to reach out to Ross for Solar’s access to the server, as Solar might assist with the transfer if Rob is overwhelmed. Rob explained the benefits of using S3 for large datasets, including infinite read bandwidth and built-in backup capabilities. Solar inquired about adapting the current retrieval algorithms for S3 URLs, and Rob assured that while some refactoring might be necessary, the team has in-house expertise to find a solution that works for GMT users.

Data Storage Migration to S3

The team discussed transitioning data storage from a Y server to S3 buckets, with Joaquim explaining that while the code currently finds tile names automatically, they would need to change the default server address when switching to S3, which would require updating GMT to a new version. Solar and Joaquim clarified that while the data preparation process would need to be adjusted to ensure files are ready for direct upload to S3, the actual upload process would be handled by automatic scripts on mirrors, similar to their current system.

Data Transfer and S3 Migration

The team discussed issues with data transfer and CI test failures due to network problems, with Joaquim explaining that moving to S3 could help address these issues by providing better speed and stability. Solar agreed to focus on understanding where remote dataset calls are happening in GMT and explore the ramifications of pulling datasets from an S3 bucket. The group also touched on the upcoming GMT Summit plans, which Rob is overseeing, and the potential release of GMT 6.7 in late April.

GMT Release 6.7 Planning Meeting

The team discussed plans for GMT release 6.7, targeting the end of April. Joaquim explained he would prepare the Windows installer, fixing issues with the large library and GPS breather models. The group also considered Dangan Tian’s suggestion to separate the documentation from the codebase, which Federico supported. Solar introduced herself as the Earthscope lead developer for GMT, managing the project and supporting its growth. The conversation ended with Joaquim and Christian joining the call, after experiencing time zone synchronization issues.

GMT and MB System Collaboration

Christian, a long-time GMT user and collaborator of David from MB System, shared his journey with the software and his recent initiatives. He has been developing a water column tool for MB System, which is available on GitHub, and is interested in creating a Python wrapper for MB System to increase its user base, particularly among R and Python users in Germany. Christian expressed interest in collaborating with the GMT group and mentioned potential funding opportunities through proposals in Germany and the EU to support GMT and MB System development. Joaquim, while closer to retirement, expressed his continued commitment to the project, and Christian emphasized the goal of establishing a collaborative group to maintain and develop MB System for the long term.

GMT Classic and Modern Modes

The discussion focused on GMT’s classic and modern modes, with Joaquim explaining that modern mode is built on top of classic mode and doesn’t eliminate PostScript dependency. Joaquim expressed concerns about GMT’s reliance on PostScript, noting that it’s becoming less widely used and could pose future risks if the software that runs PostScript were to be discontinued. Christian inquired about plans to phase out classic mode, but Joaquim clarified that while nothing is permanent, classic mode remains more powerful for certain tasks and cannot be completely separated from modern mode.

Documentation and Code Repository Separation

Joaquim and Solar discussed the potential separation of documentation and code repositories. Joaquim explained that while the documentation and code are currently integrated, they could be separated with minimal dependencies, as the code only needs to know where to find the documentation files during the build process. Solar suggested investigating the dependencies further to determine if a separation would be feasible, though they did not decide on a timeline for such a change.

GMT Expansion and Funding Discussion

The meeting focused on expanding GMT participation and exploring funding opportunities. Christian shared insights from a German government agency (BGR) that heavily uses GMT, suggesting potential collaboration for future proposals. Joaquim emphasized the importance of creating a wish list for GMT development and suggested using AI tools for documentation. The group discussed the need to attract new developers, particularly in C programming, and agreed to explore scheduling meetings at times more convenient for Asian participants. Solar provided an update on GMT’s transition to Earthscope management and its current funding status. The conversation ended with plans to include Yvonne in future meetings and to further discuss the new GMT version in the next session.

14

Hmmm, interesting: none of the transcripts explicitly mentions fixing bugs.

15

I’ve not invested in bug fixing with AI, but a fix new features were done with it’s help, the spheres in PostScript and color gradient in polygons are an excellent example, and this one was done entirely by Claude (with my help). Add a report of p-value to the regress's -Fp option. by joa-quim · Pull Request #8919 · GenericMappingTools/gmt · GitHub

16

Hmmm, could that thing possibly write or keep the list of changes? Maybe starting from the missing changelog for gmt 6.6.0?

17

I’m sure it can (was planing to use it next time). But regarding the 6.6.0 changelog, I must have done something wrong during the release because, although very short, I did one. It just doesn’t show up don’t know why.

18

Here is the changelog for 6.6.

19

Thanks Federico,

I happen to particularly like that extremely short changelog.