chore: minor reformat and rewrite

src/thesis/3-tools.typ

@@ -189,7 +189,7 @@ test scenarios described in this chapter are executed through GitHub Actions
 #cite(<ghActions>, form: "normal"). GitHub Actions is an automation platform
 that enables #gls("CICD"), allowing builds to be performed, tested, and deployed
 across various machines and architectures directly from GitHub repositories
-  #cite(<9978190>,form:"normal").
+#cite(<9978190>,form:"normal").
 
 Our testing environments supports three distinct architectures:
 

src/thesis/4-conclusion.typ

@@ -196,16 +196,17 @@ Guix has been an interesting tool to evaluate. While the learning curve is
 steeper than Docker, the benefits are significant. I appreciated the strict and
 declarative approach to package management, which aligns well with the
 reproducibility goals. The idea of using an existing general purpose language
-for declaring packages and configurations is a powerful idea. The community is
-small but active, however since no proprietary tools are packaged, it can be a
-challenge for users to find the software they need. There are workarounds
-existing but it is not advertised by the Guix community which tend to focus and
-adhere to the free software philosophy #cite(<fsfwebsite>, form: "normal"). The
-performance of Guix is great, since no containerization is involved, the
-software runs natively on the system and accessing storage and network is a
-breeze. Guix extensively uses `git` #cite(<git>, form: "normal") for fetching
-packages and configurations, and the information displayed to the user while
-running it is very clean and clear.
+for declaring packages and configurations is a powerful idea. The Guix community
+is small but active. While no proprietary tools are packaged, which can be a
+challenge for users to find the software they need, this has no impact on Guix's
+reproducibility capabilities and features. There are workarounds existing but it
+is not advertised by the Guix community which tend to focus and adhere to the
+free software philosophy #cite(<fsfwebsite>, form: "normal"). The performance of
+Guix is great, since no containerization is involved, the software runs natively
+on the system and accessing storage and network is a breeze. Guix extensively
+uses `git` #cite(<git>, form: "normal") for fetching packages and
+configurations, and the information displayed to the user while running it is
+very clean and clear.
 
 Nix has been the most interesting approaches to evaluate, technically but also
 politically. The learning curve is steep, but the benefits are significant.
@@ -627,11 +628,11 @@ Educating students in best practices is crucial for fostering a culture of
 reproducibility. We can draw a compelling parallel with mathematics, where
 reproducibility is inherently embedded. Just as mathematical proofs and
 solutions can be independently verified by anyone following the same steps and
-logic, reproducibility in #gls("SE") aims for the same level of
-transparency and verifiability. This bridge underscores that reproducibility is
-a desirable trait and a foundational principle that should be rigorously applied
-in computational research. For example, a professor might provide students with
-the necessary data, software, and materials for a specific course through fully
+logic, reproducibility in #gls("SE") aims for the same level of transparency
+and verifiability. This bridge underscores that reproducibility is a desirable
+trait and a foundational principle that should be rigorously applied in
+computational research. For example, a professor might provide students with the
+necessary data, software, and materials for a specific course through fully
 reproducible methods and tools. This could involve using open-source software,
 version control systems, and detailed documentation to ensure that students can
 reproduce the outcome in any space and any time.