FreeCAD News

The fol­low­ing text is a guest post from Turan Furkan Topak report­ing on their trip to the GSOC Sum­mit. Many thanks for tak­ing the time to write this report. /Jo

When I applied to be a men­tor for GSoC, my inten­tion was actu­al­ly just to men­tor in my own area of inter­est. That was how the process worked any­way: you were, in a way, a can­di­date can­di­date. But because we didn’t have many men­tors, I end­ed up serv­ing as a sec­ond men­tor on all the projects. And there wasn’t even a project relat­ed to doc­u­men­ta­tion, which is my area of interest.

Then, because of some polit­i­cal events, mes­sages start­ed going around that the men­tor sum­mit shouldn’t be held in the US. To be hon­est, I wasn’t real­ly pay­ing much atten­tion at the time. But in the end the deci­sion was made and the sum­mit was moved to Munich, Ger­many. I still wasn’t that invest­ed, since I knew I would have to go through a dif­fi­cult visa process and I wasn’t sure if it would be worth it. I told Yorik, the orga­ni­za­tion admin, that I would like to go if no one else want­ed to attend, and no one else stepped for­ward. By chance, I came across an oppor­tu­ni­ty that would have me in Sofia dur­ing the sum­mit and give me a long-term visa. After that, there was noth­ing left to stop me from going.

Of course the first and most sen­si­ble option that comes to mind is to fly. But when you have the chance to real­ly see, or rather pass through, most of Europe, why just fly over it? So yes, I bought bus tick­ets for both direc­tions. Bul­gar­ia, then Ser­bia, Croa­t­ia, Slove­nia, Aus­tria and final­ly Ger­many. A 23-hour trip. What could pos­si­bly go wrong?

Nat­u­ral­ly, going on such a long trip means prepar­ing pro­vi­sions. At first I was think­ing of the stan­dard cheese and sala­mi, but then I thought: why not cook some chick­en, shred it and use that instead of sala­mi? And since there would also be the return trip, I need­ed to make plen­ty. I added some toma­to and may­on­naise for a bit of col­or. We also had mush­rooms at home and they def­i­nite­ly wouldn’t be in bet­ter shape when I came back, so I cooked them togeth­er with the chick­en and made six sand­wich­es, each enough for a full meal.

I went a bit ear­ly and start­ed wait­ing for my bus, and there I met an Aus­tri­an pho­tog­ra­ph­er who was wait­ing with us. He was also a moun­taineer and had very heavy equip­ment with him. We chat­ted a lot dur­ing the breaks on the jour­ney. Because we set off at night, my plan to wit­ness the scenery of the Balka­ns had to wait. Since we were pass­ing through non-EU coun­tries, we had pass­port checks at the bor­ders, and dur­ing those I talked with the Aus­tri­an pho­tog­ra­ph­er. Even­tu­al­ly the sun rose and I could final­ly reach my goal. Where I come from, if you trav­el by bus you most­ly see end­less steppe: small shrubs, earth in shades between yel­low and orange, and rocks that you can imag­ine as any­thing. That has its own charm, of course, but the Balka­ns are some­thing else entire­ly. Vast land­scapes con­tain­ing every shade of green stretched out to hills that, with the morn­ing mist hang­ing above them, looked like mighty moun­tains whose sum­mits touched the clouds. In some val­leys there were towns, and the mix of old and new hous­es made the view even more inter­est­ing. I won­dered who owned that old wood­en house and how it was still stand­ing. Smoke from the chim­neys blend­ed into the mist on the hills. It was a pro­ces­sion of views you can’t real­ly pho­to­graph, only see.

The esti­mat­ed arrival time was 20:00. I was wor­ried about whether I would get there quick­ly enough to catch din­ner. That wor­ry was resolved short­ly after­wards. Our bus broke down near the Croa­t­ian bor­der. My trav­el com­pan­ion was head­ing some­where else, so the replace­ment bus for his route arrived quick­ly, but I had to wait for the next bus to Munich, which would come sev­en hours later.

There must have been anoth­er bus that had bro­ken down where ours did, because I came across a note left behind by six friends who had done a Europe tour in 2022, and it made me smile a lit­tle. I won­dered how long they had had to wait.

Even­tu­al­ly our bus arrived and I got on, but by then the scenery had retreat­ed and left only dark­ness behind. My arrival time no longer looked rea­son­able either. I fell asleep while try­ing to fig­ure out how I was going to get to the hotel. Some­times wait­ing tires you out even more, espe­cial­ly wait­ing this long.

We arrived in Munich at 3 a.m. and I woke up. Unfor­tu­nate­ly, if I want­ed to use pub­lic trans­port, I would have had to wait a long time. One advan­tage of trav­el by bus is that ter­mi­nals are not as far out of the city as air­ports, so the hotel was only 7 km away, along a route going through the city. I wasn’t going to have time to explore Munich, and I want­ed to hear the sound of this huge city. Cities have their own sounds, and you usu­al­ly hear them on long night walks when no one else is around. Of course my rec­om­men­da­tion is to take these night walks in sum­mer. Octo­ber looks like it should work too, but once cer­tain flute-like sounds are added to the city’s voice, you may start to think it’s not such a good idea. After a while my nav­i­ga­tion app start­ed act­ing up and send­ing me into dark parks with no lights, so I turned it off and tried to find the way myself. I almost got lost near a place where they were clear­ing away the remains of a tram acci­dent, but even­tu­al­ly I found my path. I reached the hotel around 5 and checked in. As I passed by the GSoC boards, I saw that the day would start at 7:30. I had missed enough already and didn’t want to miss any more.

I got up ear­ly, picked up my badge and the oth­er rel­e­vant things, and then went to the restau­rant for break­fast. I filled a plate from the buf­fet and sat at an emp­ty table. Lat­er, a girl who had come with her moth­er because of visa issues sat down at my table with her. We talked most­ly about how hard it is to get visas. She was quite talk­a­tive, and as the end of break­fast time approached, her plate was still full. At 9:15 the Google team gave the open­ing talk and then the orga­ni­za­tions’ short talks began. They main­ly talked about one of that year’s GSoC projects. I was also going to give a talk, but in the after­noon ses­sion. Among the speak­ers there was some­one I was sure shared my native lan­guage. Just as Dante, even when he went to hell, was look­ing for some­one to whom he could say “Si”, I, in this beau­ti­ful place, want­ed to approach some­one to whom I could say “yes” in my own lan­guage and talk about GSoC, so after the meet­ing end­ed we chat­ted a bit. He said it was the first year their orga­ni­za­tion had par­tic­i­pat­ed and that they had two stu­dents; one had dis­ap­peared dur­ing the process, and the oth­er, although good at com­mu­ni­cat­ing, wasn’t actu­al­ly pro­gress­ing on the project.

After the talks, the ses­sions began. First of all, this sum­mit is a con­fer­ence with­out a con­fer­ence. You only real­ly under­stand what that means when you get there. In short, you write down an inter­est­ing top­ic you want to talk about and stick it on the board in the time slot of the rel­e­vant room, and who­ev­er wants to talk about that top­ic comes to that room. Among the first ses­sions there was also a GSoC Feed­back ses­sion, but I was more inter­est­ed in a ses­sion on which tools are used to com­mu­ni­cate with­in com­mu­ni­ties, so I went to that. The rea­son was that there had recent­ly been a dis­cus­sion inside FreeCAD about some com­mu­ni­ca­tion tools, and I was curi­ous about the sit­u­a­tion in oth­er orga­ni­za­tions. How­ev­er, the per­son who had pro­posed the ses­sion didn’t show up, and since it was the first ses­sion, it most­ly turned into intro­duc­tions. It wasn’t a very crowd­ed ses­sion either. We also talked about the tools that were being used. The sit­u­a­tions the atten­dees described weren’t very sim­i­lar to the ones we expe­ri­ence. First of all, FreeCAD is much more inter­twined with its user base, and that user base spans gen­er­a­tions. This is prob­a­bly because we have a large user com­mu­ni­ty. In this ses­sion I espe­cial­ly appre­ci­at­ed how inter­twined the FreeCAD com­mu­ni­ty is. It was great to meet new peo­ple and most­ly lis­ten to their com­mu­ni­ca­tion adven­tures in GSoC. One par­tic­i­pant even gave me a GSoC 2025 bracelet that his wife had made.

For the next slot I joined the ses­sion on the future of soft­ware engi­neer­ing with AI. I approach the whole AI sit­u­a­tion as philo­soph­i­cal­ly as I do prag­mat­i­cal­ly. In my view, “real” AI should rule the entire world, but what is cur­rent­ly being sold to us is an illu­sion that is very far from being an intel­li­gence. You could say it’s smarter than many peo­ple, but the accu­rate state­ment is that it looks smarter than many peo­ple. In this ses­sion they also men­tioned AI’s impact on the train­ing of juniors, and I insist­ed that its impact on the envi­ron­ment be added to the dis­cus­sion. Per­son­al­ly, I don’t see AI itself as harm­ful to the envi­ron­ment; the tru­ly harm­ful thing is the peo­ple try­ing to turn it into mon­ey. And then there was the top­ic that took up the most space: copy­right. Copy­right was orig­i­nal­ly intro­duced as a monop­oly to encour­age invent­ing new things. Like every monop­oly, it became cor­rupt and con­cen­trat­ed in cer­tain hands. But what can be sub­ject to copy­right? In the­o­ry, every­thing, but once com­mer­cial ele­ments are involved, it becomes restrict­ed. Here’s an inter­est­ing bit of infor­ma­tion: humans are born able to walk, but every baby has to invent crawl­ing for them­selves and opti­mize it. If that’s the case, can crawl­ing be copy­right­ed, since it’s some­thing invent­ed? That’s why copy­right law says that an inven­tion must be some­thing that could not eas­i­ly be cre­at­ed by a per­son skilled in the art. Of course that prob­a­bly makes every issue quite com­pli­cat­ed. AI makes it even more inter­est­ing. Can some­thing dis­cov­ered by an enti­ty that isn’t even an expert in the field be sub­ject to copy­right? Which leads us to this: can what we cur­rent­ly call AI real­ly dis­cov­er any­thing? Or is it just a slight­ly more advanced ver­sion of the “Ctrl+C” buttons?

And final­ly it was time for lunch. It was a won­der­ful buf­fet of East Asian food, prob­a­bly my favorite menu of the whole sum­mit. But because of the nature of the sum­mit, you bare­ly even have time to eat. While you’re meet­ing peo­ple and talk­ing about their orga­ni­za­tions, you sud­den­ly real­ize the next hour has arrived before you’ve fin­ished your plate. Dur­ing lunch I talked about our GSoC process­es with a British veg­an gen­tle­man from an envi­ron­ment-relat­ed orga­ni­za­tion. After­wards we went up to the ball­room for the group pho­to, squeezed togeth­er quite a bit, and the pic­ture was taken.

After that, unfor­tu­nate­ly, the two ses­sions most rel­e­vant to FreeCAD were at the same time. One was a 2D/3D geom­e­try ses­sion and the oth­er was an open source hard­ware ses­sion. Sad­ly, I had to choose one. Per­son­al­ly, I hate choos­ing. But I picked the open source hard­ware ses­sion. Lat­er I learned that the 2D/3D geom­e­try ses­sion had turned into a full-on CAD meet­up and FreeCAD was the only one miss­ing; they even asked me why I hadn’t come. I was hon­est­ly sad I couldn’t be in two places at once. Since I’ve briefly men­tioned the ses­sion I missed, let me talk about the one I attended.

With FreeCAD 1.0 we gained a start screen fea­ture, and we actu­al­ly want to refresh those images with every release, but we need to put in some work for that. I also think FreeCAD should be more inter­twined with the open source hard­ware com­mu­ni­ty and be part of a more mutu­al­is­tic ecosys­tem. As the writer, there’s some­thing else I want to note. I remem­ber who every­one I spoke to at the sum­mit was and which insti­tu­tion they were from, but I thought it would be bet­ter not to men­tion that. How­ev­er, here I’ll be talk­ing about some projects, and in at least some cas­es it wouldn’t make sense to hide the names. The ses­sion was orga­nized by FOSSASIA, and they asked me what Realthun­der is doing these days; one of them said they had met him at a con­fer­ence before, and I replied that he hasn’t been very active in FreeCAD late­ly. They had brought some cool badge dis­plays with them. We start­ed by talk­ing in gen­er­al about open source hard­ware: licens­ing, main­tain­abil­i­ty, mod­u­lar­i­ty, needs, safe­ty and repro­ducibil­i­ty. Then every­one began, one by one, to show what they are cur­rent­ly work­ing on. Robert Kaye talked about his holo­gram fan project and said he had last used Bar­tendo, the auto­mat­ic cock­tail-mak­ing machine that was everyone’s favorite, at a house par­ty. Every­one asked why such a won­der­ful device wasn’t on every street cor­ner. Mr. Kaye was about to answer, but let me cut in a lit­tle. When you go to a bar, is what you are real­ly buy­ing the drink? In most cas­es the per­son there sim­ply trans­fers the liq­uid from the bot­tle to the glass. In cas­es like cock­tails, they mix the things in those bot­tles. But what you’re actu­al­ly pay­ing for is the same thing you pay for at the the­ater: the right to watch a per­for­mance, and the fact that this per­for­mance is being done for you. It’s the price of enjoy­ing, for a lit­tle while, the same sit­u­a­tion as a king. Back to our top­ic: Mr. Kaye said the device is expen­sive to pro­duce, and that when they offered it to bars, the bar­tenders were against any­thing that might replace them. Then Till Kamp­peter, the man who makes your print­er work if you use Lin­ux, talked about the Open Print­er project. He said it is some­thing extreme­ly easy to under­stand and repair, and that they would soon be open­ing it for fund­ing. After that, Deep­ak Kha­tri showed BeagleConnect’s new Zep­to board. He list­ed a long and impres­sive set of fea­tures for such a small board and said its cost was very low. When I told him it sound­ed almost like he was talk­ing about the Holy Grail, he laughed. He also showed some cool projects from his own start­up, Upside Down Lab, like dri­ving an RC car with brain pow­er and mak­ing music with mus­cle move­ments. When we spoke again lat­er, I learned that he uses FreeCAD while doing these things. When it was my turn, unfor­tu­nate­ly I didn’t have a project of my own, but I said that if any­one in the room had a project made with FreeCAD that they thought would inter­est peo­ple, we could put it on the start screen or on our site. They real­ly liked the ani­mat­ed con­tent on the site, and a lot of peo­ple seemed inter­est­ed. But time flew by and the ses­sion came to an end.

I actu­al­ly thought the light­ning talks were next, and my talk was among the first, so I went to the ball­room, but I had mixed up the times. The talks were an hour lat­er; at that moment they were dis­cussing the GSoC+AI sur­vey in the ball­room. As in every AI con­ver­sa­tion, the top­ic broad­ened and turned to the orga­ni­za­tions’ poli­cies. Some­one asked who among those present had an orga­ni­za­tion that banned the use of AI. A rep­re­sen­ta­tive said they had com­plete­ly banned it in their orga­ni­za­tion. Some­one else, quite rea­son­ably, asked how they made sure whether some­thing involved AI or not, and the answer was that these two things were unre­lat­ed. And that’s real­ly how it is: detect­ing whether some­thing was done by AI has noth­ing to do with whether you accept AI as a mat­ter of pol­i­cy. If AI con­tributes some­thing so good that it can’t be told apart, is that real­ly AI? But there is a prob­lem here: no rule should push peo­ple towards cheat­ing. The pur­pose of rules is to be excep­tion­al mea­sures that pre­vent bad sit­u­a­tions. Free­dom is the default. Here I also shared the AI pol­i­cy we use in FreeCAD: “Con­tri­bu­tions must meet exist­ing qual­i­ty stan­dards. Raw AI out­put is not accept­ed under any cir­cum­stances. AI may be used only as an assis­tive tool; in all cas­es, the result­ing con­tent must be reviewed, val­i­dat­ed, and jus­ti­fi­able by the con­trib­u­tor. The con­trib­u­tor should be able to explain design and code deci­sions, answer review­ers’ ques­tions, and ensure that AI use does not waste review­ers’ time dur­ing review. That being said, the use of AI is not rec­om­mend­ed under any cir­cum­stances or in any man­ner.” As you can see, our pol­i­cy is tuned to pre­vent sit­u­a­tions we don’t want, not to enforce our per­son­al opin­ions, but for the health of the project. And regard­ing copy­right, which is anoth­er prob­lem area: this is not just an issue with AI, so reg­u­lat­ing it only on the basis of AI would be wrong; a human can cre­ate the same prob­lems here. Our pol­i­cy was very well received; some­one even said the oth­ers present should steal it. In fact, it may be wrong to talk about the prob­lems AI cur­rent­ly cre­ates, because in the future it will cause prob­lems dif­fer­ent from the ones it caus­es now, and we may even­tu­al­ly reach a point where we can’t even dis­tin­guish it from a human. I believe our pol­i­cy is future-proof with that in mind.

Then the light­ning talks came, and I was sixth in line. I was hon­est­ly quite ner­vous on stage. While I’m at it, I’d like to once again thank Theo-vt, Cap­tain and Chi­rag for being such great stu­dents. And of course Kadet, Pieter­hi­j­ma, Chennes and Cad­di­a­log for being such great men­tors. And of course Yorik, who wasn’t a men­tor this year but kept the orga­ni­za­tion run­ning. In my short talk I want­ed to men­tion every­one and all the amaz­ing things they did, but unfor­tu­nate­ly I couldn’t fit it all in. From what I heard and learned talk­ing with men­tors from oth­er orga­ni­za­tions, I can say that we were one of the orga­ni­za­tions with the high­est suc­cess rate this year. After the pre­sen­ta­tions, an email arrived say­ing that speak­ers could pick some gifts, and I chose a few things from the list.

Since there was time until din­ner, I want­ed to chat with oth­er men­tors and see the famous choco­late room. From what I learned, the choco­late room is a GSoC sum­mit tra­di­tion: peo­ple from all over the world bring choco­late from where they come from so oth­ers can taste it. I tried many of them; I found some real favorites, and there was also a lot of very high-per­cent­age dark choco­late. There I talked with anoth­er men­tor about what states real­ly are. Should states be an aver­age of the peo­ple who make them up, or should they try to move those peo­ple for­ward? We talked a bit about lan­guage, about how you can nev­er quite escape your moth­er tongue. In the end we rec­om­mend­ed books to each oth­er. I sug­gest­ed Diana Wynne Jones’s Howl’s Mov­ing Cas­tle tril­o­gy. All three books are very good, but I like the film more; it’s a won­der­ful pot­pour­ri of all three.

And of course, din­ner even­tu­al­ly arrived. They were mak­ing secret prepa­ra­tions in the ball­room that they were try­ing to hide from every­one. If I had tak­en a moment to think about it, I could prob­a­bly have guessed what they were doing, but I didn’t have the time even for that. It was Octo­ber and we were in Ger­many, so of course they were prepar­ing some­thing with an Okto­ber­fest theme. It was the only GSoC sum­mit with a cow. Despite the music play­ing in the back­ground, I kept chat­ting with the men­tors next to me. After din­ner we said good­bye to Stephanie, who was leav­ing ear­ly in the morn­ing. And as you could guess from the pres­ence of a sound sys­tem, karaoke began. One of the fun­ni­est things, in my opin­ion, was that they used YouTube for it and we had to sit through long ads. Then some peo­ple stepped up and found a way to block the ads. You would expect some­thing like that from a group like this. In keep­ing with the spir­it of the day, I sang John Lennon’s “Imag­ine”. I def­i­nite­ly did not secret­ly type “eas­i­est karaoke songs to sing” into the inter­net. Some­one even con­grat­u­lat­ed me on my choice. After a while group per­for­mances and even a list formed, but it was get­ting late, and we end­ed the night with the hall manager’s per­for­mance of “Clos­ing Time”.

I got up at half past sev­en for break­fast and, of course, spent the time talk­ing with some men­tors again. I had to check out dur­ing the day, and since there wasn’t any­thing in the next ses­sions that inter­est­ed me that much, I went up to my room, packed, then checked out and left my things in storage.

Next I attend­ed Google’s own talk on doc­u­men­ta­tion. It dis­cussed an analy­sis done for their sis­ter project, Google Sea­son of Docs, which had been dis­con­tin­ued the pre­vi­ous year. They had hired some­one to ana­lyze the prob­lems there, and that per­son had pre­pared arche­types for doc­u­men­ta­tion con­trib­u­tors. The stick­ers they made for those arche­types were great, but I can’t say the actu­al study was. First of all, I can say that most orga­ni­za­tions at the meet­ing didn’t dis­tin­guish between devel­op­er and user doc­u­men­ta­tion. Maybe the orga­ni­za­tions there don’t need such a dis­tinc­tion; I don’t know. But even so, the study didn’t touch on this in any oth­er con­text either. I also learned that trans­la­tion sup­port is usu­al­ly han­dled sep­a­rate­ly. For FreeCAD, trans­la­tion sup­port is offi­cial­ly pro­vid­ed from a sin­gle cen­tral place. My view for user doc­u­men­ta­tion is that any­one who reads it should be able to con­tribute. What needs to be done is to make con­tribut­ing as easy as pos­si­ble and encour­age peo­ple to do it. Anoth­er issue is syn­chro­niza­tion between doc­u­men­ta­tion and code, that is, mak­ing sure that a new­ly added fea­ture is also added to the doc­u­men­ta­tion. I think this could be done with some kind of GitHub work­flow. But that won’t be hap­pen­ing any time soon.

The next ses­sion ref­er­enced a meme about open source tak­ing up a huge place in pub­lic infra­struc­ture. Nat­u­ral­ly, the top­ic was how, despite being so impor­tant, that infra­struc­ture is not sup­port­ed. How much do the insti­tu­tions that use open source tools in such crit­i­cal places actu­al­ly give back? Unfor­tu­nate­ly, very lit­tle. And do we real­ly want to be that guy in Nebras­ka, or are we that resilient? It was a great ses­sion on funding.

The next slot was unfor­tu­nate­ly only a half-hour ses­sion. I went to the Vin­tage Com­put­ing ses­sion. We talked about some cool machines: the ones they cur­rent­ly have and the ones that had passed through their hands, and the dif­fi­cul­ties involved. It was a fun conversation.

And then it was lunchtime. There, a men­tor from a 2D-relat­ed orga­ni­za­tion asked me why I hadn’t come to the 2D/3D geom­e­try ses­sion, and I told him about the oth­er ses­sion. Then we talked about how orga­ni­za­tions work. I men­tioned how use­ful our meet­ings are and described our meet­ing struc­ture: a week­ly merge and fund­ing meet­ing and a biweek­ly devel­op­er meet­ing. He said they might try some of those things.

Lat­er I joined a ses­sion where we shared strange sto­ries we had expe­ri­enced or heard about AI. My favorite sto­ry was about spam copy­right notices: at first they sim­ply said there was no such thing, and lat­er they start­ed reply­ing with long AI-gen­er­at­ed peti­tions designed to keep the oth­er side busy for a long time. Before send­ing their next spam, the oth­er side had to wade through pages of mean­ing­less text. Anoth­er sto­ry was about an AI used to pro­mote cars in a gallery in Cana­da that was tricked into sell­ing a car for one dol­lar. We also talked about the curi­ous kinds of fun AI brings into our lives.

In the last ses­sion I attend­ed Robert Kaye’s meet­ing about the Euro­pean Open Source Academy’s awards. He talked about how it’s an orga­ni­za­tion aimed at increas­ing the vis­i­bil­i­ty and fund­ing of open source in Europe, and how open source should be a full-time job, even though most peo­ple there still work in jobs unre­lat­ed to open source. He also talked about the academy’s nom­i­na­tion process. You can nom­i­nate any­one you like through the academy’s website.

And with that, all the ses­sions were over. I bought dried veg­eta­bles, fruit and nuts to eat on the way. While wait­ing for the bus I chat­ted with oth­er men­tors and tried to find the oth­er CAD men­tors. But what I was actu­al­ly wait­ing for wasn’t the bus stand­ing at the plat­form; I had already begun wait­ing for the next GSoC to come, for the day I’d get to expe­ri­ence this mag­i­cal atmos­phere and process all over again.