It made me sad.
I’m working on a project involving 911 data and it’s made me do quite a bit of thinking. The short story on this is it’s not in Tennessee. I’m taking data in it’s raw form and shoving it into that state’s NG911 format. So I have about half of it “pre-written” with existing scripts.
I was lucky in Tennessee as the Standard was written around a database and left the user to “fix” the data by whatever means you wanted. You could edit the data in Autocad (shudder) if you wanted and as long as it was loaded into the format the state was expecting you were good. So I made it all work in QGIS/PostGIS. It works well enough now that I’m down to quality checks and whether these run automatically or by the user. One runs every night. One runs when needed. In both cases nothing may happen because you can ignore both options.
This pile of data has been way different. Being stupid lead me down one path and I failed. They have a toolset in Arcpy to help you find errors and send the data to the state. I had to fall back to the toolset. Which meant standing up a Windows 11 VM and running ArcMap. At one point I had an error aaaand had to fall back to editing one of the Gigantic Arcpy (4000+ lines of code) to get it to work. So the last part of this has to run through ArcGIS to succeed. ArcMap is officially deprecated. I have Python 2.X code I’m digging through on Windows 11. Ugh….BUT….I’m slowly getting the SQL code built out to do this all the way I want to do it….which is annoying because this could turn into a one off project. So I’m not killing myself getting the sql code done.
At the Georgia Geospatial Conference I gave a rambling presentation revolving around “How to convert a proprietary project into an open source one” when really I’m thinking I should give something along the “The Rambling Guide to introducing your org to open source GIS”….or something. I haven’t quite figured out how to explain what I want to explain.
Twice a year at some random conference I have someone come up to me and go Company XYZ is killing me with invoices for software. Can I go open source? Can we switch it all out next Tuesday at 9 AM? Sort of a “drop in and replace” everything with QGIS, QGIS Server, QGIS Online, QGIS Database etc. I always answer that request with a question of “Are you insane?” and “no”.
So here’s what I tell people and few like the explanation BUT I’m probably going to re-write my presentation at FOSS4GNA with this. This is way over simplified.
- Find a small project no one cares about. Maybe they minimally care about it. Ask people. Get people to go “Yeah it’s not critical”.
- Install PostGIS. Install QGIS.
- Manage the entire thing using QGIS/PostGIS.
- At this point No one may care. You may find that literally “no one cares” and maybe you trash it. Maybe you continue.
- At some point someone may ask and bring another equally boring project to you and you shove it into PostGIS.
- Perhaps you need to share it and you bring Geoserver into the Mix.
- Again – you may get to this part and everyone is having fits and you may need to abandon the effort….OR….you can point out this hasn’t cost any money except for a VM or a minimal Cloud install.
- Then believe it or not you’re using a mixed Environment. Things that work well in QGIS are working well. Hopefully things that work well with your other software keep not being a problem.
- By this point you’ve joined a lot of community support sites (slack/discourse/email) and you’re minimally talking to the community going “Hey I did a thing”.
- You find someone else did the same thing or close to it. Now there are two of you. Maybe more.
- This takes probably a year. Maybe a bit more.
This current project I’m on may not see the full shift to opensource because they’ve built the entire thing around ArcMap. I had one giant failure. I’m now having one partial success. Maybe by project #3 that’s a 100% success. With any luck we have a process not built off dead software by the end of this.
Anyway – Go do some good out there.
