Update a few sentences in paper.md

Author: amontoison

paper/paper.bib

@@ -48,12 +48,24 @@ @article{jonasson2020
   number    = {1},
   pages     = {1--20},
   year      = {2020},
+  doi       = {10.1145/3382191},
   publisher = {ACM New York, NY, USA}
 }
 
 @article{giles2008,
   title     = {An extended collection of matrix derivative results for forward and reverse mode automatic differentiation},
   author    = {Giles, Mike},
   year      = {2008},
-  publisher = {Unspecified}
+}
+
+@inproceedings{TAF,
+  title        = {{Applying TAF to generate efficient derivative code of Fortran 77-95 programs}},
+  author       = {Giering, Ralf and Kaminski, Thomas},
+  booktitle    = {{PAMM: Proceedings in Applied Mathematics and Mechanics}},
+  volume       = {2},
+  number       = {1},
+  pages        = {54--57},
+  year         = {2003},
+  doi          = {10.1002/pamm.200310014},
+  organization = {Wiley Online Library}
 }

paper/paper.md

@@ -27,29 +27,30 @@ bibliography: paper.bib
 
 # Summary
 
-`diffblas` is a library that provides algorithmically differentiated [@griewank2008] BLAS routines from their reference implementations in [LAPACK](https://github.com/Reference-LAPACK/lapack) on GitHub using the automatic differentiation tool Tapenade [@tapenade].
+`diffblas` is a library that provides BLAS routines algorithmically differentiated using algorithmic differentiation principles [@griewank2008] from their reference implementations in [LAPACK](https://github.com/Reference-LAPACK/lapack) on GitHub using the automatic differentiation tool Tapenade [@tapenade].
 It supports four modes: forward (`_d`), vector forward (`_dv`), reverse (`_b`), and vector reverse (`_bv`).
 
 In addition to differentiating the standard Fortran-style `BLAS` interface, `diffblas` also provides differentiated `CBLAS` routines, facilitating interoperability with C and other languages.
-Its API mirrors BLAS/CBLAS, with additional arguments specifying differentiation variables, making it straightforward to integrate into existing workflows.
+Its API mirrors BLAS / CBLAS, with additional arguments specifying differentiation variables, making integration into existing workflows straightforward.
 
 `diffblas` calls the underlying standard `BLAS `implementation, and is agnostic to the backend (OpenBLAS, BLIS, MKL, Apple Accelerate), ensuring both performance and portability.
+Precompiled artifacts are available on GitHub to simplify installation and usage across different programming environments.
 
-By providing accurate and efficient derivatives of linear algebra operations, `diffblas` facilitates gradient-based optimization, sensitivity analysis, and a wide range of scientific computing applications
+By providing efficient and accurate derivatives of linear algebra operations, `diffblas` facilitates gradient-based optimization, sensitivity analysis, and derivative-based workflows in scientific computing.
 
 # Statement of need
 
-Linear algebra routines such as those in LAPACK are fundamental to scientific computing, optimization, and machine learning.
-However, they do not provide derivatives, which are often required for gradient-based algorithms.
-Current approaches either rely on hand-coded derivatives or generic automatic differentiation applied to high-level code, which can be inefficient or error-prone.
+Linear algebra routines such as those in LAPACK are widely used in scientific computing, optimization, and machine learning. However, they do not provide derivatives, which are often required for gradient-based algorithms.
+
+Existing approaches rely on hand-coded derivatives or generic automatic differentiation applied to high-level code, which can be inefficient or error-prone [@jonasson2020].
+
 `diffblas` addresses this gap by providing algorithmically differentiated BLAS routines directly from reference LAPACK implementations and following relevant differntiation rules [@giles2008].
-This enables accurate and efficient computation of derivatives while preserving compatibility with existing BLAS-based codes.
 
 # State of the field
 
-Automatic differentiation (AD) tools such as Tapenade [@tapenade], ADOL-C [@ADOLC], or Taf provide general mechanisms to compute derivatives of code.
-However, applying AD naively to low-level BLAS or LAPACK routines can be inefficient due to loops, memory layout, and caching issues [@jonasson2020].
-Specialized libraries like diffblas that generate differentiated routines directly from reference implementations combine the reliability of LAPACK with the efficiency of AD, bridging a gap in current scientific computing workflows.
+Automatic source-to-source differentiation tools, such as Tapenade [@tapenade], ADOL-C [@ADOLC], or TAF [@TAF], provide general mechanisms to compute derivatives of code.
+However, naively applying AD to low-level BLAS or LAPACK routines can be inefficient due to loop structures, memory layout, and caching constraints [@jonasson2020].
+`diffblas` addresses these challenges by generating differentiated routines directly from reference LAPACK implementations, combining the reliability and performance of LAPACK with the efficiency of algorithmic differentiation.
 
 # Research impact statement