Moving a personal academic website. A checklist.

(revised Aug 21, 2026)

After 25 years of hosting at ITU web servers (at the late http://www.itu.dk/people/wasowski/) my website moved to wasowski.dukla.ch. This coincided with a sudden decision of ITU’s IT Department to discontinue the service (rants suppressed). Now the website runs on my own hardware in my living room. Be my guest!

The situation is also a side-effect of Danish universities becoming increasingly corporate and top-down managed. We reached a point, when hosting a faculty website at the university infrastructure is not acceptable for so-called “legal reasons.” The only approved communication is apparently the one of the professionals in the communication departments. The expertise achieved through producing years of peer reviewed research is apparently not to be trusted. Is this a good development? Let’s ponder the question a bit. … eh, better not :)

Legally, take note, this means that, this website has nothing to do with my employment. All opinions, rights, and blames, are mine.

Redirect the Internet to your new location

The most interesting aspect of this exercise, was trying to redirect the main pointers of the Internet to my new location, to help others discover it. I was surprised with how many places deserve a check. I summarized the checklist, so others do not need to rediscover it from scratch.

1. Academic profiles The online profiles you own (if you are an academic you definitely should claim at least a few of the top of this list, if you have not done this yet).
   • Google Scholar (mine). You can edit the link yourself, if you have claimed your profile. Sadly the main Google search engine ignores it.
   • ORCID (mine). Edit the link yourself. It is publicly visible.
   • DBLP (mine). Requires an email to dblp support and within a few weeks the nice people in Dagstuhl will fix it. The link is publicly visible, from the tiny drop down menu in the top. You do care for DBLP, if your field is computer science.
   • ACM Author Profile (mine). If you publish in computing, you probably have one, and you do not even know. Most likely, as it was in my case, it was automatically generated and contained a lot of incorrect information. It allows you to create a public link to your personal website, and correct some of the incorrect data ACM has on you. The edits are moderated, so it takes some days. This profile is linked from all of your papers in the ACM Digital Library.
   • IEEE Author Profile (mine). Similar to ACM, but the tech seems to be a few years behind. As of today, not first class link, but you can add it as clear text to your bio. Like with ACM, the edits are moderated, so it takes some days to activate. Also like with ACM, this allows you to correct spelling of your name, add alternative names, etc. This is a useful thing for folks who have non-latin alphabet names (commonly get messed up), have multi-part names (commonly get messed up), or have changed names.
   • conf.researchr.org (mine). You have this profile if you publish, review, or co-organize events that use this service for hosting conference websites. In software engineering and programming languages this is fairly common. The website has a good traction, and shows a lot of your activities (I would say it is more useful than Elsevier’s Pure profiles). It allows you to link to your personal website, and probably has a stale link if you moved.
   • There might be other, non-researchr.org-hosted conferences, professional associations, journal editorial boards, and similar, where you have service roles. Typically you can update them. You know what they are.
   • Springer and Elsevier do not maintain useful author profiles (in the sense that they can host your link).
   • arxiv.org has a homepage link in their meta-data. I couldn’t figure out how this affects visibility. Probably doesn’t—they never show it, but it could be used to disambiguate your identity in the future, so you might want to fix it.
   • openreview.net (mine). This is a public link and you can change it yourself.
2. Code and data forges
   • In my case both codeberg.org and github.com both host personal profiles. You might be using more code and data serving accounts (hugging face?). I closed my personal workspace on bitbucket.org, and I think Atlassian does not host personal profiles, but check, if this is relevant for you.
3. Social media
   • Fediverse instance profile (mine), or any other social media profile you care for. If you are not on fediverse yet, let me use this opportunity to invite you to become a member of the most social social-media on the planet. (NB. Incidentally my profile is running on the same machine as this website).
   • LinkedIn (mine) and other corporate media you care for. The website link is pretty well hidden in the profile settings on LinkedIn, but you can change it yourself. Not many people will get to it, but it probably matters for Bing, and the main profile with the link seems to be visible for non-members, so it might be indexable by other search engines.)
   • Scientific Social Media: ResearchGate, Academia.edu. Each has a profile of you, and while I do not care for them (they seem more dodgy than corporate social media), them pointing to your old website is harming you. Claim the profiles and update the link.
4. University services
   • Faculty roster (mine). Our faculty roster is configured (still) to allow linking to external websites. The roster page is a publicly accessible link so totally worth it.
   • Learning management system (mine, requires an ITU account). ITU’s moodle-based study management system has a faculty profile page with a personal website link visible to potential MSc and BSc students. You can edit the profile and change the link yourself.
   • Local Github Enterprise installation (mine, requires ITU account), gitlab, Forgejo, or whatever your school uses. I share code and data with students via this route, so some students might learn something about me this way.
5. Other websites you control or semi-control, including websites of friends, coauthors, and family.
   • Research group website, or other organizational unit websites in your school that might allow links to personal pages.
   • Past and current research projects often contain a list of PIs with links. For instance I could still update https://remaro.eu/index.php/supervisors/.
   • Supplement websites for papers and books. In my case our book website
   • Websites of other academics that could be bothered to ask. A quick online search for my name showed an odd link from a colleague’s website. This was worth an email (they are better off not linking to a service that just produces a beautifully professional 503).
6. Documents
   • CV. The PDF file with your CV that you keep regularly updated. You do, right?
   • Email signature. You have probably forgotten that it exists. I kept propagating a dead link for several weeks :)
   • Template for your slides for course lectures and conference talks
   • Letterhead template if you use one.
7. Profiles in Research Funding Agencies, so that you are mapped to the right research proposals as a reviewer. I only remembered two, where I am registered. Both had a link to my website as meta-data.
   • European Commission Expert profile.
   • European Science Foundation profile (sadly, I could not modify the field, when I tried.)
8. Dungeons of the Internet and of some libraries, not so important, but while we are at it…
   • Wikidata. This is something I did not know about before. Even if you do not have a wikipedia page (I don’t), there is likely a wikidata entry for your name (mine). Wikidata is a part of the knowledge graph infrastructure of the Internet. Supposedly, wikidata was to blame for Google presenting my photo next to a bio of a pianist that died years ago. My profile was a fusion of three persons with the same name. Worth cleaning up, linking to your website, and checking other services it links to (mostly bibliographic services of libraries and scientific institutions in my case).
9. Search engine consoles. This is important.
   • Google. Set up an account for your new website at Google Search Console. Tell the search engine about your new address. Create a sitemap.xml and register it in the console (Hugo creates the map for free). In my experience, Google is reluctant to listen. They are picky about your HTML, redirect behaviours, uptime and such. It is useful to revisit its console regularly to eliminate the errors it points to. Google down ranks pages with errors. In general, they seem to throttle the crawl budget for new domains to near-nothing, so getting their attention will take time.
   • Bing Roughly the same task as for Google, but use Bing Webmaster Tools. In my experience, Bing is pretty fast to pick up new pages. Their index is also used by “independent” search engines (Duck Duck Go, Ecosia, Qwant, etc.).
   • Others. I have not done anything for Baidu and Yandex, but some referrals came anyway.

This was a lot of work, so I hope I do not need to move again for the next 25 years, hopefully never. After all this, your new website is properly a member of the academic internet graph. And search engines will pick it up. In my case, Google has been reluctant, so now I have disappeared from their web for many months. Bing and independent search engines picked up the new website pretty fast, though.

There are rumours that choosing an established domain, such as *.github.io, or a global level domain (.com, .org, .edu, .net) helps positioning with Google. I ruled out the former, because I did not want to depend on another corporation for my online presence. The latter was tempting, but at the point of migration, I had owned dukla.ch for some years, so I simply decided to use a resource I already had.

Technicalities

For a short period I was able to operate a redirect for all subpages moved, but this did not last long. There are two ways to do this: a server level redirect on modern HTTP servers, or page stubs that do HTML-level redirect (yes our server was ancient). Early summer 2026, the ITU Apache installation has been powered off, so these redirects no longer work.

For nerds, the website is served from a small, 10+ years old, Intel NUC machine running Arch Linux, which also arranges a lot of other small chores for me. Since the machine lives behind a NAT, it is proxied through a VPS on public IP net via a small installation of Caddy and WireGuard. The website is static, generated with Hugo using the PaperMod theme. So far I am not using nodejs, or any tracking libraries (I hope!). No JavaScript beyond what the theme uses for the light/dark toggle.